4 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
6 * This program is free software; you can redistribute it and/or modify
7 * it under the terms of the GNU General Public License version 2 only,
8 * as published by the Free Software Foundation.
10 * This program is distributed in the hope that it will be useful, but
11 * WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 * General Public License version 2 for more details (a copy is included
14 * in the LICENSE file that accompanied this code).
16 * You should have received a copy of the GNU General Public License
17 * version 2 along with this program; If not, see
18 * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
20 * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
21 * CA 95054 USA or visit www.sun.com if you need additional information or
27 * Copyright (c) 2002, 2010, Oracle and/or its affiliates. All rights reserved.
28 * Use is subject to license terms.
30 * Copyright (c) 2011, 2012, Whamcloud, Inc.
33 * This file is part of Lustre, http://www.lustre.org/
34 * Lustre is a trademark of Sun Microsystems, Inc.
37 #define DEBUG_SUBSYSTEM S_OSC
39 #include <libcfs/libcfs.h>
42 # include <liblustre.h>
45 #include <lustre_dlm.h>
46 #include <lustre_net.h>
47 #include <lustre/lustre_user.h>
48 #include <obd_cksum.h>
56 #include <lustre_ha.h>
57 #include <lprocfs_status.h>
58 #include <lustre_log.h>
59 #include <lustre_debug.h>
60 #include <lustre_param.h>
61 #include "osc_internal.h"
62 #include "osc_cl_internal.h"
64 static void osc_release_ppga(struct brw_page **ppga, obd_count count);
65 static int brw_interpret(const struct lu_env *env,
66 struct ptlrpc_request *req, void *data, int rc);
67 int osc_cleanup(struct obd_device *obd);
69 /* Pack OSC object metadata for disk storage (LE byte order). */
70 static int osc_packmd(struct obd_export *exp, struct lov_mds_md **lmmp,
71 struct lov_stripe_md *lsm)
76 lmm_size = sizeof(**lmmp);
81 OBD_FREE(*lmmp, lmm_size);
87 OBD_ALLOC(*lmmp, lmm_size);
93 LASSERT(lsm->lsm_object_id);
94 LASSERT_SEQ_IS_MDT(lsm->lsm_object_seq);
95 (*lmmp)->lmm_object_id = cpu_to_le64(lsm->lsm_object_id);
96 (*lmmp)->lmm_object_seq = cpu_to_le64(lsm->lsm_object_seq);
102 /* Unpack OSC object metadata from disk storage (LE byte order). */
103 static int osc_unpackmd(struct obd_export *exp, struct lov_stripe_md **lsmp,
104 struct lov_mds_md *lmm, int lmm_bytes)
107 struct obd_import *imp = class_exp2cliimp(exp);
111 if (lmm_bytes < sizeof (*lmm)) {
112 CERROR("lov_mds_md too small: %d, need %d\n",
113 lmm_bytes, (int)sizeof(*lmm));
116 /* XXX LOV_MAGIC etc check? */
118 if (lmm->lmm_object_id == 0) {
119 CERROR("lov_mds_md: zero lmm_object_id\n");
124 lsm_size = lov_stripe_md_size(1);
128 if (*lsmp != NULL && lmm == NULL) {
129 OBD_FREE((*lsmp)->lsm_oinfo[0], sizeof(struct lov_oinfo));
130 OBD_FREE(*lsmp, lsm_size);
136 OBD_ALLOC(*lsmp, lsm_size);
139 OBD_ALLOC((*lsmp)->lsm_oinfo[0], sizeof(struct lov_oinfo));
140 if ((*lsmp)->lsm_oinfo[0] == NULL) {
141 OBD_FREE(*lsmp, lsm_size);
144 loi_init((*lsmp)->lsm_oinfo[0]);
148 /* XXX zero *lsmp? */
149 (*lsmp)->lsm_object_id = le64_to_cpu (lmm->lmm_object_id);
150 (*lsmp)->lsm_object_seq = le64_to_cpu (lmm->lmm_object_seq);
151 LASSERT((*lsmp)->lsm_object_id);
152 LASSERT_SEQ_IS_MDT((*lsmp)->lsm_object_seq);
156 (imp->imp_connect_data.ocd_connect_flags & OBD_CONNECT_MAXBYTES))
157 (*lsmp)->lsm_maxbytes = imp->imp_connect_data.ocd_maxbytes;
159 (*lsmp)->lsm_maxbytes = LUSTRE_STRIPE_MAXBYTES;
164 static inline void osc_pack_capa(struct ptlrpc_request *req,
165 struct ost_body *body, void *capa)
167 struct obd_capa *oc = (struct obd_capa *)capa;
168 struct lustre_capa *c;
173 c = req_capsule_client_get(&req->rq_pill, &RMF_CAPA1);
176 body->oa.o_valid |= OBD_MD_FLOSSCAPA;
177 DEBUG_CAPA(D_SEC, c, "pack");
180 static inline void osc_pack_req_body(struct ptlrpc_request *req,
181 struct obd_info *oinfo)
183 struct ost_body *body;
185 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
188 lustre_set_wire_obdo(&body->oa, oinfo->oi_oa);
189 osc_pack_capa(req, body, oinfo->oi_capa);
192 static inline void osc_set_capa_size(struct ptlrpc_request *req,
193 const struct req_msg_field *field,
197 req_capsule_set_size(&req->rq_pill, field, RCL_CLIENT, 0);
199 /* it is already calculated as sizeof struct obd_capa */
203 static int osc_getattr_interpret(const struct lu_env *env,
204 struct ptlrpc_request *req,
205 struct osc_async_args *aa, int rc)
207 struct ost_body *body;
213 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
215 CDEBUG(D_INODE, "mode: %o\n", body->oa.o_mode);
216 lustre_get_wire_obdo(aa->aa_oi->oi_oa, &body->oa);
218 /* This should really be sent by the OST */
219 aa->aa_oi->oi_oa->o_blksize = PTLRPC_MAX_BRW_SIZE;
220 aa->aa_oi->oi_oa->o_valid |= OBD_MD_FLBLKSZ;
222 CDEBUG(D_INFO, "can't unpack ost_body\n");
224 aa->aa_oi->oi_oa->o_valid = 0;
227 rc = aa->aa_oi->oi_cb_up(aa->aa_oi, rc);
231 static int osc_getattr_async(struct obd_export *exp, struct obd_info *oinfo,
232 struct ptlrpc_request_set *set)
234 struct ptlrpc_request *req;
235 struct osc_async_args *aa;
239 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_GETATTR);
243 osc_set_capa_size(req, &RMF_CAPA1, oinfo->oi_capa);
244 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_GETATTR);
246 ptlrpc_request_free(req);
250 osc_pack_req_body(req, oinfo);
252 ptlrpc_request_set_replen(req);
253 req->rq_interpret_reply = (ptlrpc_interpterer_t)osc_getattr_interpret;
255 CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
256 aa = ptlrpc_req_async_args(req);
259 ptlrpc_set_add_req(set, req);
263 static int osc_getattr(const struct lu_env *env, struct obd_export *exp,
264 struct obd_info *oinfo)
266 struct ptlrpc_request *req;
267 struct ost_body *body;
271 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_GETATTR);
275 osc_set_capa_size(req, &RMF_CAPA1, oinfo->oi_capa);
276 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_GETATTR);
278 ptlrpc_request_free(req);
282 osc_pack_req_body(req, oinfo);
284 ptlrpc_request_set_replen(req);
286 rc = ptlrpc_queue_wait(req);
290 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
292 GOTO(out, rc = -EPROTO);
294 CDEBUG(D_INODE, "mode: %o\n", body->oa.o_mode);
295 lustre_get_wire_obdo(oinfo->oi_oa, &body->oa);
297 /* This should really be sent by the OST */
298 oinfo->oi_oa->o_blksize = PTLRPC_MAX_BRW_SIZE;
299 oinfo->oi_oa->o_valid |= OBD_MD_FLBLKSZ;
303 ptlrpc_req_finished(req);
307 static int osc_setattr(const struct lu_env *env, struct obd_export *exp,
308 struct obd_info *oinfo, struct obd_trans_info *oti)
310 struct ptlrpc_request *req;
311 struct ost_body *body;
315 LASSERT(oinfo->oi_oa->o_valid & OBD_MD_FLGROUP);
317 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_SETATTR);
321 osc_set_capa_size(req, &RMF_CAPA1, oinfo->oi_capa);
322 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_SETATTR);
324 ptlrpc_request_free(req);
328 osc_pack_req_body(req, oinfo);
330 ptlrpc_request_set_replen(req);
332 rc = ptlrpc_queue_wait(req);
336 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
338 GOTO(out, rc = -EPROTO);
340 lustre_get_wire_obdo(oinfo->oi_oa, &body->oa);
344 ptlrpc_req_finished(req);
348 static int osc_setattr_interpret(const struct lu_env *env,
349 struct ptlrpc_request *req,
350 struct osc_setattr_args *sa, int rc)
352 struct ost_body *body;
358 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
360 GOTO(out, rc = -EPROTO);
362 lustre_get_wire_obdo(sa->sa_oa, &body->oa);
364 rc = sa->sa_upcall(sa->sa_cookie, rc);
368 int osc_setattr_async_base(struct obd_export *exp, struct obd_info *oinfo,
369 struct obd_trans_info *oti,
370 obd_enqueue_update_f upcall, void *cookie,
371 struct ptlrpc_request_set *rqset)
373 struct ptlrpc_request *req;
374 struct osc_setattr_args *sa;
378 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_SETATTR);
382 osc_set_capa_size(req, &RMF_CAPA1, oinfo->oi_capa);
383 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_SETATTR);
385 ptlrpc_request_free(req);
389 if (oti && oinfo->oi_oa->o_valid & OBD_MD_FLCOOKIE)
390 oinfo->oi_oa->o_lcookie = *oti->oti_logcookies;
392 osc_pack_req_body(req, oinfo);
394 ptlrpc_request_set_replen(req);
396 /* do mds to ost setattr asynchronously */
398 /* Do not wait for response. */
399 ptlrpcd_add_req(req, PDL_POLICY_ROUND, -1);
401 req->rq_interpret_reply =
402 (ptlrpc_interpterer_t)osc_setattr_interpret;
404 CLASSERT (sizeof(*sa) <= sizeof(req->rq_async_args));
405 sa = ptlrpc_req_async_args(req);
406 sa->sa_oa = oinfo->oi_oa;
407 sa->sa_upcall = upcall;
408 sa->sa_cookie = cookie;
410 if (rqset == PTLRPCD_SET)
411 ptlrpcd_add_req(req, PDL_POLICY_ROUND, -1);
413 ptlrpc_set_add_req(rqset, req);
419 static int osc_setattr_async(struct obd_export *exp, struct obd_info *oinfo,
420 struct obd_trans_info *oti,
421 struct ptlrpc_request_set *rqset)
423 return osc_setattr_async_base(exp, oinfo, oti,
424 oinfo->oi_cb_up, oinfo, rqset);
427 int osc_real_create(struct obd_export *exp, struct obdo *oa,
428 struct lov_stripe_md **ea, struct obd_trans_info *oti)
430 struct ptlrpc_request *req;
431 struct ost_body *body;
432 struct lov_stripe_md *lsm;
441 rc = obd_alloc_memmd(exp, &lsm);
446 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_CREATE);
448 GOTO(out, rc = -ENOMEM);
450 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_CREATE);
452 ptlrpc_request_free(req);
456 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
458 lustre_set_wire_obdo(&body->oa, oa);
460 ptlrpc_request_set_replen(req);
462 if ((oa->o_valid & OBD_MD_FLFLAGS) &&
463 oa->o_flags == OBD_FL_DELORPHAN) {
465 "delorphan from OST integration");
466 /* Don't resend the delorphan req */
467 req->rq_no_resend = req->rq_no_delay = 1;
470 rc = ptlrpc_queue_wait(req);
474 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
476 GOTO(out_req, rc = -EPROTO);
478 lustre_get_wire_obdo(oa, &body->oa);
480 /* This should really be sent by the OST */
481 oa->o_blksize = PTLRPC_MAX_BRW_SIZE;
482 oa->o_valid |= OBD_MD_FLBLKSZ;
484 /* XXX LOV STACKING: the lsm that is passed to us from LOV does not
485 * have valid lsm_oinfo data structs, so don't go touching that.
486 * This needs to be fixed in a big way.
488 lsm->lsm_object_id = oa->o_id;
489 lsm->lsm_object_seq = oa->o_seq;
493 oti->oti_transno = lustre_msg_get_transno(req->rq_repmsg);
495 if (oa->o_valid & OBD_MD_FLCOOKIE) {
496 if (!oti->oti_logcookies)
497 oti_alloc_cookies(oti, 1);
498 *oti->oti_logcookies = oa->o_lcookie;
502 CDEBUG(D_HA, "transno: "LPD64"\n",
503 lustre_msg_get_transno(req->rq_repmsg));
505 ptlrpc_req_finished(req);
508 obd_free_memmd(exp, &lsm);
512 int osc_punch_base(struct obd_export *exp, struct obd_info *oinfo,
513 obd_enqueue_update_f upcall, void *cookie,
514 struct ptlrpc_request_set *rqset)
516 struct ptlrpc_request *req;
517 struct osc_setattr_args *sa;
518 struct ost_body *body;
522 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_PUNCH);
526 osc_set_capa_size(req, &RMF_CAPA1, oinfo->oi_capa);
527 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_PUNCH);
529 ptlrpc_request_free(req);
532 req->rq_request_portal = OST_IO_PORTAL; /* bug 7198 */
533 ptlrpc_at_set_req_timeout(req);
535 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
537 lustre_set_wire_obdo(&body->oa, oinfo->oi_oa);
538 osc_pack_capa(req, body, oinfo->oi_capa);
540 ptlrpc_request_set_replen(req);
542 req->rq_interpret_reply = (ptlrpc_interpterer_t)osc_setattr_interpret;
543 CLASSERT (sizeof(*sa) <= sizeof(req->rq_async_args));
544 sa = ptlrpc_req_async_args(req);
545 sa->sa_oa = oinfo->oi_oa;
546 sa->sa_upcall = upcall;
547 sa->sa_cookie = cookie;
548 if (rqset == PTLRPCD_SET)
549 ptlrpcd_add_req(req, PDL_POLICY_ROUND, -1);
551 ptlrpc_set_add_req(rqset, req);
556 static int osc_punch(const struct lu_env *env, struct obd_export *exp,
557 struct obd_info *oinfo, struct obd_trans_info *oti,
558 struct ptlrpc_request_set *rqset)
560 oinfo->oi_oa->o_size = oinfo->oi_policy.l_extent.start;
561 oinfo->oi_oa->o_blocks = oinfo->oi_policy.l_extent.end;
562 oinfo->oi_oa->o_valid |= OBD_MD_FLSIZE | OBD_MD_FLBLOCKS;
563 return osc_punch_base(exp, oinfo,
564 oinfo->oi_cb_up, oinfo, rqset);
567 static int osc_sync_interpret(const struct lu_env *env,
568 struct ptlrpc_request *req,
571 struct osc_fsync_args *fa = arg;
572 struct ost_body *body;
578 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
580 CERROR ("can't unpack ost_body\n");
581 GOTO(out, rc = -EPROTO);
584 *fa->fa_oi->oi_oa = body->oa;
586 rc = fa->fa_upcall(fa->fa_cookie, rc);
590 int osc_sync_base(struct obd_export *exp, struct obd_info *oinfo,
591 obd_enqueue_update_f upcall, void *cookie,
592 struct ptlrpc_request_set *rqset)
594 struct ptlrpc_request *req;
595 struct ost_body *body;
596 struct osc_fsync_args *fa;
600 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_SYNC);
604 osc_set_capa_size(req, &RMF_CAPA1, oinfo->oi_capa);
605 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_SYNC);
607 ptlrpc_request_free(req);
611 /* overload the size and blocks fields in the oa with start/end */
612 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
614 lustre_set_wire_obdo(&body->oa, oinfo->oi_oa);
615 osc_pack_capa(req, body, oinfo->oi_capa);
617 ptlrpc_request_set_replen(req);
618 req->rq_interpret_reply = osc_sync_interpret;
620 CLASSERT(sizeof(*fa) <= sizeof(req->rq_async_args));
621 fa = ptlrpc_req_async_args(req);
623 fa->fa_upcall = upcall;
624 fa->fa_cookie = cookie;
626 if (rqset == PTLRPCD_SET)
627 ptlrpcd_add_req(req, PDL_POLICY_ROUND, -1);
629 ptlrpc_set_add_req(rqset, req);
634 static int osc_sync(const struct lu_env *env, struct obd_export *exp,
635 struct obd_info *oinfo, obd_size start, obd_size end,
636 struct ptlrpc_request_set *set)
641 CDEBUG(D_INFO, "oa NULL\n");
645 oinfo->oi_oa->o_size = start;
646 oinfo->oi_oa->o_blocks = end;
647 oinfo->oi_oa->o_valid |= (OBD_MD_FLSIZE | OBD_MD_FLBLOCKS);
649 RETURN(osc_sync_base(exp, oinfo, oinfo->oi_cb_up, oinfo, set));
652 /* Find and cancel locally locks matched by @mode in the resource found by
653 * @objid. Found locks are added into @cancel list. Returns the amount of
654 * locks added to @cancels list. */
655 static int osc_resource_get_unused(struct obd_export *exp, struct obdo *oa,
657 ldlm_mode_t mode, int lock_flags)
659 struct ldlm_namespace *ns = exp->exp_obd->obd_namespace;
660 struct ldlm_res_id res_id;
661 struct ldlm_resource *res;
665 osc_build_res_name(oa->o_id, oa->o_seq, &res_id);
666 res = ldlm_resource_get(ns, NULL, &res_id, 0, 0);
670 LDLM_RESOURCE_ADDREF(res);
671 count = ldlm_cancel_resource_local(res, cancels, NULL, mode,
672 lock_flags, 0, NULL);
673 LDLM_RESOURCE_DELREF(res);
674 ldlm_resource_putref(res);
678 static int osc_destroy_interpret(const struct lu_env *env,
679 struct ptlrpc_request *req, void *data,
682 struct client_obd *cli = &req->rq_import->imp_obd->u.cli;
684 cfs_atomic_dec(&cli->cl_destroy_in_flight);
685 cfs_waitq_signal(&cli->cl_destroy_waitq);
689 static int osc_can_send_destroy(struct client_obd *cli)
691 if (cfs_atomic_inc_return(&cli->cl_destroy_in_flight) <=
692 cli->cl_max_rpcs_in_flight) {
693 /* The destroy request can be sent */
696 if (cfs_atomic_dec_return(&cli->cl_destroy_in_flight) <
697 cli->cl_max_rpcs_in_flight) {
699 * The counter has been modified between the two atomic
702 cfs_waitq_signal(&cli->cl_destroy_waitq);
707 /* Destroy requests can be async always on the client, and we don't even really
708 * care about the return code since the client cannot do anything at all about
710 * When the MDS is unlinking a filename, it saves the file objects into a
711 * recovery llog, and these object records are cancelled when the OST reports
712 * they were destroyed and sync'd to disk (i.e. transaction committed).
713 * If the client dies, or the OST is down when the object should be destroyed,
714 * the records are not cancelled, and when the OST reconnects to the MDS next,
715 * it will retrieve the llog unlink logs and then sends the log cancellation
716 * cookies to the MDS after committing destroy transactions. */
717 static int osc_destroy(const struct lu_env *env, struct obd_export *exp,
718 struct obdo *oa, struct lov_stripe_md *ea,
719 struct obd_trans_info *oti, struct obd_export *md_export,
722 struct client_obd *cli = &exp->exp_obd->u.cli;
723 struct ptlrpc_request *req;
724 struct ost_body *body;
725 CFS_LIST_HEAD(cancels);
730 CDEBUG(D_INFO, "oa NULL\n");
734 count = osc_resource_get_unused(exp, oa, &cancels, LCK_PW,
735 LDLM_FL_DISCARD_DATA);
737 req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_DESTROY);
739 ldlm_lock_list_put(&cancels, l_bl_ast, count);
743 osc_set_capa_size(req, &RMF_CAPA1, (struct obd_capa *)capa);
744 rc = ldlm_prep_elc_req(exp, req, LUSTRE_OST_VERSION, OST_DESTROY,
747 ptlrpc_request_free(req);
751 req->rq_request_portal = OST_IO_PORTAL; /* bug 7198 */
752 ptlrpc_at_set_req_timeout(req);
754 if (oti != NULL && oa->o_valid & OBD_MD_FLCOOKIE)
755 oa->o_lcookie = *oti->oti_logcookies;
756 body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
758 lustre_set_wire_obdo(&body->oa, oa);
760 osc_pack_capa(req, body, (struct obd_capa *)capa);
761 ptlrpc_request_set_replen(req);
763 /* don't throttle destroy RPCs for the MDT */
764 if (!(cli->cl_import->imp_connect_flags_orig & OBD_CONNECT_MDS)) {
765 req->rq_interpret_reply = osc_destroy_interpret;
766 if (!osc_can_send_destroy(cli)) {
767 struct l_wait_info lwi = LWI_INTR(LWI_ON_SIGNAL_NOOP,
771 * Wait until the number of on-going destroy RPCs drops
772 * under max_rpc_in_flight
774 l_wait_event_exclusive(cli->cl_destroy_waitq,
775 osc_can_send_destroy(cli), &lwi);
779 /* Do not wait for response */
780 ptlrpcd_add_req(req, PDL_POLICY_ROUND, -1);
784 static void osc_announce_cached(struct client_obd *cli, struct obdo *oa,
787 obd_flag bits = OBD_MD_FLBLOCKS|OBD_MD_FLGRANT;
789 LASSERT(!(oa->o_valid & bits));
792 client_obd_list_lock(&cli->cl_loi_list_lock);
793 oa->o_dirty = cli->cl_dirty;
794 if (cli->cl_dirty - cli->cl_dirty_transit > cli->cl_dirty_max) {
795 CERROR("dirty %lu - %lu > dirty_max %lu\n",
796 cli->cl_dirty, cli->cl_dirty_transit, cli->cl_dirty_max);
798 } else if (cfs_atomic_read(&obd_dirty_pages) -
799 cfs_atomic_read(&obd_dirty_transit_pages) >
800 obd_max_dirty_pages + 1){
801 /* The cfs_atomic_read() allowing the cfs_atomic_inc() are
802 * not covered by a lock thus they may safely race and trip
803 * this CERROR() unless we add in a small fudge factor (+1). */
804 CERROR("dirty %d - %d > system dirty_max %d\n",
805 cfs_atomic_read(&obd_dirty_pages),
806 cfs_atomic_read(&obd_dirty_transit_pages),
807 obd_max_dirty_pages);
809 } else if (cli->cl_dirty_max - cli->cl_dirty > 0x7fffffff) {
810 CERROR("dirty %lu - dirty_max %lu too big???\n",
811 cli->cl_dirty, cli->cl_dirty_max);
814 long max_in_flight = (cli->cl_max_pages_per_rpc << CFS_PAGE_SHIFT)*
815 (cli->cl_max_rpcs_in_flight + 1);
816 oa->o_undirty = max(cli->cl_dirty_max, max_in_flight);
818 oa->o_grant = cli->cl_avail_grant;
819 oa->o_dropped = cli->cl_lost_grant;
820 cli->cl_lost_grant = 0;
821 client_obd_list_unlock(&cli->cl_loi_list_lock);
822 CDEBUG(D_CACHE,"dirty: "LPU64" undirty: %u dropped %u grant: "LPU64"\n",
823 oa->o_dirty, oa->o_undirty, oa->o_dropped, oa->o_grant);
827 void osc_update_next_shrink(struct client_obd *cli)
829 cli->cl_next_shrink_grant =
830 cfs_time_shift(cli->cl_grant_shrink_interval);
831 CDEBUG(D_CACHE, "next time %ld to shrink grant \n",
832 cli->cl_next_shrink_grant);
835 static void __osc_update_grant(struct client_obd *cli, obd_size grant)
837 client_obd_list_lock(&cli->cl_loi_list_lock);
838 cli->cl_avail_grant += grant;
839 client_obd_list_unlock(&cli->cl_loi_list_lock);
842 static void osc_update_grant(struct client_obd *cli, struct ost_body *body)
844 if (body->oa.o_valid & OBD_MD_FLGRANT) {
845 CDEBUG(D_CACHE, "got "LPU64" extra grant\n", body->oa.o_grant);
846 __osc_update_grant(cli, body->oa.o_grant);
850 static int osc_set_info_async(const struct lu_env *env, struct obd_export *exp,
851 obd_count keylen, void *key, obd_count vallen,
852 void *val, struct ptlrpc_request_set *set);
854 static int osc_shrink_grant_interpret(const struct lu_env *env,
855 struct ptlrpc_request *req,
858 struct client_obd *cli = &req->rq_import->imp_obd->u.cli;
859 struct obdo *oa = ((struct osc_grant_args *)aa)->aa_oa;
860 struct ost_body *body;
863 __osc_update_grant(cli, oa->o_grant);
867 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
869 osc_update_grant(cli, body);
875 static void osc_shrink_grant_local(struct client_obd *cli, struct obdo *oa)
877 client_obd_list_lock(&cli->cl_loi_list_lock);
878 oa->o_grant = cli->cl_avail_grant / 4;
879 cli->cl_avail_grant -= oa->o_grant;
880 client_obd_list_unlock(&cli->cl_loi_list_lock);
881 if (!(oa->o_valid & OBD_MD_FLFLAGS)) {
882 oa->o_valid |= OBD_MD_FLFLAGS;
885 oa->o_flags |= OBD_FL_SHRINK_GRANT;
886 osc_update_next_shrink(cli);
889 /* Shrink the current grant, either from some large amount to enough for a
890 * full set of in-flight RPCs, or if we have already shrunk to that limit
891 * then to enough for a single RPC. This avoids keeping more grant than
892 * needed, and avoids shrinking the grant piecemeal. */
893 static int osc_shrink_grant(struct client_obd *cli)
895 long target = (cli->cl_max_rpcs_in_flight + 1) *
896 cli->cl_max_pages_per_rpc;
898 client_obd_list_lock(&cli->cl_loi_list_lock);
899 if (cli->cl_avail_grant <= target)
900 target = cli->cl_max_pages_per_rpc;
901 client_obd_list_unlock(&cli->cl_loi_list_lock);
903 return osc_shrink_grant_to_target(cli, target);
906 int osc_shrink_grant_to_target(struct client_obd *cli, long target)
909 struct ost_body *body;
912 client_obd_list_lock(&cli->cl_loi_list_lock);
913 /* Don't shrink if we are already above or below the desired limit
914 * We don't want to shrink below a single RPC, as that will negatively
915 * impact block allocation and long-term performance. */
916 if (target < cli->cl_max_pages_per_rpc)
917 target = cli->cl_max_pages_per_rpc;
919 if (target >= cli->cl_avail_grant) {
920 client_obd_list_unlock(&cli->cl_loi_list_lock);
923 client_obd_list_unlock(&cli->cl_loi_list_lock);
929 osc_announce_cached(cli, &body->oa, 0);
931 client_obd_list_lock(&cli->cl_loi_list_lock);
932 body->oa.o_grant = cli->cl_avail_grant - target;
933 cli->cl_avail_grant = target;
934 client_obd_list_unlock(&cli->cl_loi_list_lock);
935 if (!(body->oa.o_valid & OBD_MD_FLFLAGS)) {
936 body->oa.o_valid |= OBD_MD_FLFLAGS;
937 body->oa.o_flags = 0;
939 body->oa.o_flags |= OBD_FL_SHRINK_GRANT;
940 osc_update_next_shrink(cli);
942 rc = osc_set_info_async(NULL, cli->cl_import->imp_obd->obd_self_export,
943 sizeof(KEY_GRANT_SHRINK), KEY_GRANT_SHRINK,
944 sizeof(*body), body, NULL);
946 __osc_update_grant(cli, body->oa.o_grant);
951 #define GRANT_SHRINK_LIMIT PTLRPC_MAX_BRW_SIZE
952 static int osc_should_shrink_grant(struct client_obd *client)
954 cfs_time_t time = cfs_time_current();
955 cfs_time_t next_shrink = client->cl_next_shrink_grant;
957 if ((client->cl_import->imp_connect_data.ocd_connect_flags &
958 OBD_CONNECT_GRANT_SHRINK) == 0)
961 if (cfs_time_aftereq(time, next_shrink - 5 * CFS_TICK)) {
962 if (client->cl_import->imp_state == LUSTRE_IMP_FULL &&
963 client->cl_avail_grant > GRANT_SHRINK_LIMIT)
966 osc_update_next_shrink(client);
971 static int osc_grant_shrink_grant_cb(struct timeout_item *item, void *data)
973 struct client_obd *client;
975 cfs_list_for_each_entry(client, &item->ti_obd_list,
976 cl_grant_shrink_list) {
977 if (osc_should_shrink_grant(client))
978 osc_shrink_grant(client);
983 static int osc_add_shrink_grant(struct client_obd *client)
987 rc = ptlrpc_add_timeout_client(client->cl_grant_shrink_interval,
989 osc_grant_shrink_grant_cb, NULL,
990 &client->cl_grant_shrink_list);
992 CERROR("add grant client %s error %d\n",
993 client->cl_import->imp_obd->obd_name, rc);
996 CDEBUG(D_CACHE, "add grant client %s \n",
997 client->cl_import->imp_obd->obd_name);
998 osc_update_next_shrink(client);
1002 static int osc_del_shrink_grant(struct client_obd *client)
1004 return ptlrpc_del_timeout_client(&client->cl_grant_shrink_list,
1008 static void osc_init_grant(struct client_obd *cli, struct obd_connect_data *ocd)
1011 * ocd_grant is the total grant amount we're expect to hold: if we've
1012 * been evicted, it's the new avail_grant amount, cl_dirty will drop
1013 * to 0 as inflight RPCs fail out; otherwise, it's avail_grant + dirty.
1015 * race is tolerable here: if we're evicted, but imp_state already
1016 * left EVICTED state, then cl_dirty must be 0 already.
1018 client_obd_list_lock(&cli->cl_loi_list_lock);
1019 if (cli->cl_import->imp_state == LUSTRE_IMP_EVICTED)
1020 cli->cl_avail_grant = ocd->ocd_grant;
1022 cli->cl_avail_grant = ocd->ocd_grant - cli->cl_dirty;
1024 if (cli->cl_avail_grant < 0) {
1025 CWARN("%s: available grant < 0, the OSS is probably not running"
1026 " with patch from bug20278 (%ld) \n",
1027 cli->cl_import->imp_obd->obd_name, cli->cl_avail_grant);
1028 /* workaround for 1.6 servers which do not have
1029 * the patch from bug20278 */
1030 cli->cl_avail_grant = ocd->ocd_grant;
1033 client_obd_list_unlock(&cli->cl_loi_list_lock);
1035 CDEBUG(D_CACHE, "%s, setting cl_avail_grant: %ld cl_lost_grant: %ld \n",
1036 cli->cl_import->imp_obd->obd_name,
1037 cli->cl_avail_grant, cli->cl_lost_grant);
1039 if (ocd->ocd_connect_flags & OBD_CONNECT_GRANT_SHRINK &&
1040 cfs_list_empty(&cli->cl_grant_shrink_list))
1041 osc_add_shrink_grant(cli);
1044 /* We assume that the reason this OSC got a short read is because it read
1045 * beyond the end of a stripe file; i.e. lustre is reading a sparse file
1046 * via the LOV, and it _knows_ it's reading inside the file, it's just that
1047 * this stripe never got written at or beyond this stripe offset yet. */
1048 static void handle_short_read(int nob_read, obd_count page_count,
1049 struct brw_page **pga)
1054 /* skip bytes read OK */
1055 while (nob_read > 0) {
1056 LASSERT (page_count > 0);
1058 if (pga[i]->count > nob_read) {
1059 /* EOF inside this page */
1060 ptr = cfs_kmap(pga[i]->pg) +
1061 (pga[i]->off & ~CFS_PAGE_MASK);
1062 memset(ptr + nob_read, 0, pga[i]->count - nob_read);
1063 cfs_kunmap(pga[i]->pg);
1069 nob_read -= pga[i]->count;
1074 /* zero remaining pages */
1075 while (page_count-- > 0) {
1076 ptr = cfs_kmap(pga[i]->pg) + (pga[i]->off & ~CFS_PAGE_MASK);
1077 memset(ptr, 0, pga[i]->count);
1078 cfs_kunmap(pga[i]->pg);
1083 static int check_write_rcs(struct ptlrpc_request *req,
1084 int requested_nob, int niocount,
1085 obd_count page_count, struct brw_page **pga)
1090 remote_rcs = req_capsule_server_sized_get(&req->rq_pill, &RMF_RCS,
1091 sizeof(*remote_rcs) *
1093 if (remote_rcs == NULL) {
1094 CDEBUG(D_INFO, "Missing/short RC vector on BRW_WRITE reply\n");
1098 /* return error if any niobuf was in error */
1099 for (i = 0; i < niocount; i++) {
1100 if ((int)remote_rcs[i] < 0)
1101 return(remote_rcs[i]);
1103 if (remote_rcs[i] != 0) {
1104 CDEBUG(D_INFO, "rc[%d] invalid (%d) req %p\n",
1105 i, remote_rcs[i], req);
1110 if (req->rq_bulk->bd_nob_transferred != requested_nob) {
1111 CERROR("Unexpected # bytes transferred: %d (requested %d)\n",
1112 req->rq_bulk->bd_nob_transferred, requested_nob);
1119 static inline int can_merge_pages(struct brw_page *p1, struct brw_page *p2)
1121 if (p1->flag != p2->flag) {
1122 unsigned mask = ~(OBD_BRW_FROM_GRANT| OBD_BRW_NOCACHE|
1123 OBD_BRW_SYNC|OBD_BRW_ASYNC|OBD_BRW_NOQUOTA);
1125 /* warn if we try to combine flags that we don't know to be
1126 * safe to combine */
1127 if (unlikely((p1->flag & mask) != (p2->flag & mask))) {
1128 CWARN("Saw flags 0x%x and 0x%x in the same brw, please "
1129 "report this at http://bugs.whamcloud.com/\n",
1130 p1->flag, p2->flag);
1135 return (p1->off + p1->count == p2->off);
1138 static obd_count osc_checksum_bulk(int nob, obd_count pg_count,
1139 struct brw_page **pga, int opc,
1140 cksum_type_t cksum_type)
1144 struct cfs_crypto_hash_desc *hdesc;
1145 unsigned int bufsize;
1147 unsigned char cfs_alg = cksum_obd2cfs(cksum_type);
1149 LASSERT(pg_count > 0);
1151 hdesc = cfs_crypto_hash_init(cfs_alg, NULL, 0);
1152 if (IS_ERR(hdesc)) {
1153 CERROR("Unable to initialize checksum hash %s\n",
1154 cfs_crypto_hash_name(cfs_alg));
1155 return PTR_ERR(hdesc);
1158 while (nob > 0 && pg_count > 0) {
1159 int count = pga[i]->count > nob ? nob : pga[i]->count;
1161 /* corrupt the data before we compute the checksum, to
1162 * simulate an OST->client data error */
1163 if (i == 0 && opc == OST_READ &&
1164 OBD_FAIL_CHECK(OBD_FAIL_OSC_CHECKSUM_RECEIVE)) {
1165 unsigned char *ptr = cfs_kmap(pga[i]->pg);
1166 int off = pga[i]->off & ~CFS_PAGE_MASK;
1167 memcpy(ptr + off, "bad1", min(4, nob));
1168 cfs_kunmap(pga[i]->pg);
1170 cfs_crypto_hash_update_page(hdesc, pga[i]->pg,
1171 pga[i]->off & ~CFS_PAGE_MASK,
1173 LL_CDEBUG_PAGE(D_PAGE, pga[i]->pg, "off %d checksum %x\n",
1174 (int)(pga[i]->off & ~CFS_PAGE_MASK), cksum);
1176 nob -= pga[i]->count;
1182 err = cfs_crypto_hash_final(hdesc, (unsigned char *)&cksum, &bufsize);
1185 cfs_crypto_hash_final(hdesc, NULL, NULL);
1187 /* For sending we only compute the wrong checksum instead
1188 * of corrupting the data so it is still correct on a redo */
1189 if (opc == OST_WRITE && OBD_FAIL_CHECK(OBD_FAIL_OSC_CHECKSUM_SEND))
1195 static int osc_brw_prep_request(int cmd, struct client_obd *cli,struct obdo *oa,
1196 struct lov_stripe_md *lsm, obd_count page_count,
1197 struct brw_page **pga,
1198 struct ptlrpc_request **reqp,
1199 struct obd_capa *ocapa, int reserve,
1202 struct ptlrpc_request *req;
1203 struct ptlrpc_bulk_desc *desc;
1204 struct ost_body *body;
1205 struct obd_ioobj *ioobj;
1206 struct niobuf_remote *niobuf;
1207 int niocount, i, requested_nob, opc, rc;
1208 struct osc_brw_async_args *aa;
1209 struct req_capsule *pill;
1210 struct brw_page *pg_prev;
1213 if (OBD_FAIL_CHECK(OBD_FAIL_OSC_BRW_PREP_REQ))
1214 RETURN(-ENOMEM); /* Recoverable */
1215 if (OBD_FAIL_CHECK(OBD_FAIL_OSC_BRW_PREP_REQ2))
1216 RETURN(-EINVAL); /* Fatal */
1218 if ((cmd & OBD_BRW_WRITE) != 0) {
1220 req = ptlrpc_request_alloc_pool(cli->cl_import,
1221 cli->cl_import->imp_rq_pool,
1222 &RQF_OST_BRW_WRITE);
1225 req = ptlrpc_request_alloc(cli->cl_import, &RQF_OST_BRW_READ);
1230 for (niocount = i = 1; i < page_count; i++) {
1231 if (!can_merge_pages(pga[i - 1], pga[i]))
1235 pill = &req->rq_pill;
1236 req_capsule_set_size(pill, &RMF_OBD_IOOBJ, RCL_CLIENT,
1238 req_capsule_set_size(pill, &RMF_NIOBUF_REMOTE, RCL_CLIENT,
1239 niocount * sizeof(*niobuf));
1240 osc_set_capa_size(req, &RMF_CAPA1, ocapa);
1242 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, opc);
1244 ptlrpc_request_free(req);
1247 req->rq_request_portal = OST_IO_PORTAL; /* bug 7198 */
1248 ptlrpc_at_set_req_timeout(req);
1250 if (opc == OST_WRITE)
1251 desc = ptlrpc_prep_bulk_imp(req, page_count,
1252 BULK_GET_SOURCE, OST_BULK_PORTAL);
1254 desc = ptlrpc_prep_bulk_imp(req, page_count,
1255 BULK_PUT_SINK, OST_BULK_PORTAL);
1258 GOTO(out, rc = -ENOMEM);
1259 /* NB request now owns desc and will free it when it gets freed */
1261 body = req_capsule_client_get(pill, &RMF_OST_BODY);
1262 ioobj = req_capsule_client_get(pill, &RMF_OBD_IOOBJ);
1263 niobuf = req_capsule_client_get(pill, &RMF_NIOBUF_REMOTE);
1264 LASSERT(body != NULL && ioobj != NULL && niobuf != NULL);
1266 lustre_set_wire_obdo(&body->oa, oa);
1268 obdo_to_ioobj(oa, ioobj);
1269 ioobj->ioo_bufcnt = niocount;
1270 osc_pack_capa(req, body, ocapa);
1271 LASSERT (page_count > 0);
1273 for (requested_nob = i = 0; i < page_count; i++, niobuf++) {
1274 struct brw_page *pg = pga[i];
1275 int poff = pg->off & ~CFS_PAGE_MASK;
1277 LASSERT(pg->count > 0);
1278 /* make sure there is no gap in the middle of page array */
1279 LASSERTF(page_count == 1 ||
1280 (ergo(i == 0, poff + pg->count == CFS_PAGE_SIZE) &&
1281 ergo(i > 0 && i < page_count - 1,
1282 poff == 0 && pg->count == CFS_PAGE_SIZE) &&
1283 ergo(i == page_count - 1, poff == 0)),
1284 "i: %d/%d pg: %p off: "LPU64", count: %u\n",
1285 i, page_count, pg, pg->off, pg->count);
1287 LASSERTF(i == 0 || pg->off > pg_prev->off,
1288 "i %d p_c %u pg %p [pri %lu ind %lu] off "LPU64
1289 " prev_pg %p [pri %lu ind %lu] off "LPU64"\n",
1291 pg->pg, page_private(pg->pg), pg->pg->index, pg->off,
1292 pg_prev->pg, page_private(pg_prev->pg),
1293 pg_prev->pg->index, pg_prev->off);
1295 LASSERTF(i == 0 || pg->off > pg_prev->off,
1296 "i %d p_c %u\n", i, page_count);
1298 LASSERT((pga[0]->flag & OBD_BRW_SRVLOCK) ==
1299 (pg->flag & OBD_BRW_SRVLOCK));
1301 ptlrpc_prep_bulk_page(desc, pg->pg, poff, pg->count);
1302 requested_nob += pg->count;
1304 if (i > 0 && can_merge_pages(pg_prev, pg)) {
1306 niobuf->len += pg->count;
1308 niobuf->offset = pg->off;
1309 niobuf->len = pg->count;
1310 niobuf->flags = pg->flag;
1315 LASSERTF((void *)(niobuf - niocount) ==
1316 req_capsule_client_get(&req->rq_pill, &RMF_NIOBUF_REMOTE),
1317 "want %p - real %p\n", req_capsule_client_get(&req->rq_pill,
1318 &RMF_NIOBUF_REMOTE), (void *)(niobuf - niocount));
1320 osc_announce_cached(cli, &body->oa, opc == OST_WRITE ? requested_nob:0);
1322 if ((body->oa.o_valid & OBD_MD_FLFLAGS) == 0) {
1323 body->oa.o_valid |= OBD_MD_FLFLAGS;
1324 body->oa.o_flags = 0;
1326 body->oa.o_flags |= OBD_FL_RECOV_RESEND;
1329 if (osc_should_shrink_grant(cli))
1330 osc_shrink_grant_local(cli, &body->oa);
1332 /* size[REQ_REC_OFF] still sizeof (*body) */
1333 if (opc == OST_WRITE) {
1334 if (cli->cl_checksum &&
1335 !sptlrpc_flavor_has_bulk(&req->rq_flvr)) {
1336 /* store cl_cksum_type in a local variable since
1337 * it can be changed via lprocfs */
1338 cksum_type_t cksum_type = cli->cl_cksum_type;
1340 if ((body->oa.o_valid & OBD_MD_FLFLAGS) == 0) {
1341 oa->o_flags &= OBD_FL_LOCAL_MASK;
1342 body->oa.o_flags = 0;
1344 body->oa.o_flags |= cksum_type_pack(cksum_type);
1345 body->oa.o_valid |= OBD_MD_FLCKSUM | OBD_MD_FLFLAGS;
1346 body->oa.o_cksum = osc_checksum_bulk(requested_nob,
1350 CDEBUG(D_PAGE, "checksum at write origin: %x\n",
1352 /* save this in 'oa', too, for later checking */
1353 oa->o_valid |= OBD_MD_FLCKSUM | OBD_MD_FLFLAGS;
1354 oa->o_flags |= cksum_type_pack(cksum_type);
1356 /* clear out the checksum flag, in case this is a
1357 * resend but cl_checksum is no longer set. b=11238 */
1358 oa->o_valid &= ~OBD_MD_FLCKSUM;
1360 oa->o_cksum = body->oa.o_cksum;
1361 /* 1 RC per niobuf */
1362 req_capsule_set_size(pill, &RMF_RCS, RCL_SERVER,
1363 sizeof(__u32) * niocount);
1365 if (cli->cl_checksum &&
1366 !sptlrpc_flavor_has_bulk(&req->rq_flvr)) {
1367 if ((body->oa.o_valid & OBD_MD_FLFLAGS) == 0)
1368 body->oa.o_flags = 0;
1369 body->oa.o_flags |= cksum_type_pack(cli->cl_cksum_type);
1370 body->oa.o_valid |= OBD_MD_FLCKSUM | OBD_MD_FLFLAGS;
1373 ptlrpc_request_set_replen(req);
1375 CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
1376 aa = ptlrpc_req_async_args(req);
1378 aa->aa_requested_nob = requested_nob;
1379 aa->aa_nio_count = niocount;
1380 aa->aa_page_count = page_count;
1384 CFS_INIT_LIST_HEAD(&aa->aa_oaps);
1385 if (ocapa && reserve)
1386 aa->aa_ocapa = capa_get(ocapa);
1392 ptlrpc_req_finished(req);
1396 static int check_write_checksum(struct obdo *oa, const lnet_process_id_t *peer,
1397 __u32 client_cksum, __u32 server_cksum, int nob,
1398 obd_count page_count, struct brw_page **pga,
1399 cksum_type_t client_cksum_type)
1403 cksum_type_t cksum_type;
1405 if (server_cksum == client_cksum) {
1406 CDEBUG(D_PAGE, "checksum %x confirmed\n", client_cksum);
1410 cksum_type = cksum_type_unpack(oa->o_valid & OBD_MD_FLFLAGS ?
1412 new_cksum = osc_checksum_bulk(nob, page_count, pga, OST_WRITE,
1415 if (cksum_type != client_cksum_type)
1416 msg = "the server did not use the checksum type specified in "
1417 "the original request - likely a protocol problem";
1418 else if (new_cksum == server_cksum)
1419 msg = "changed on the client after we checksummed it - "
1420 "likely false positive due to mmap IO (bug 11742)";
1421 else if (new_cksum == client_cksum)
1422 msg = "changed in transit before arrival at OST";
1424 msg = "changed in transit AND doesn't match the original - "
1425 "likely false positive due to mmap IO (bug 11742)";
1427 LCONSOLE_ERROR_MSG(0x132, "BAD WRITE CHECKSUM: %s: from %s inode "DFID
1428 " object "LPU64"/"LPU64" extent ["LPU64"-"LPU64"]\n",
1429 msg, libcfs_nid2str(peer->nid),
1430 oa->o_valid & OBD_MD_FLFID ? oa->o_parent_seq : (__u64)0,
1431 oa->o_valid & OBD_MD_FLFID ? oa->o_parent_oid : 0,
1432 oa->o_valid & OBD_MD_FLFID ? oa->o_parent_ver : 0,
1434 oa->o_valid & OBD_MD_FLGROUP ? oa->o_seq : (__u64)0,
1436 pga[page_count-1]->off + pga[page_count-1]->count - 1);
1437 CERROR("original client csum %x (type %x), server csum %x (type %x), "
1438 "client csum now %x\n", client_cksum, client_cksum_type,
1439 server_cksum, cksum_type, new_cksum);
1443 /* Note rc enters this function as number of bytes transferred */
1444 static int osc_brw_fini_request(struct ptlrpc_request *req, int rc)
1446 struct osc_brw_async_args *aa = (void *)&req->rq_async_args;
1447 const lnet_process_id_t *peer =
1448 &req->rq_import->imp_connection->c_peer;
1449 struct client_obd *cli = aa->aa_cli;
1450 struct ost_body *body;
1451 __u32 client_cksum = 0;
1454 if (rc < 0 && rc != -EDQUOT) {
1455 DEBUG_REQ(D_INFO, req, "Failed request with rc = %d\n", rc);
1459 LASSERTF(req->rq_repmsg != NULL, "rc = %d\n", rc);
1460 body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
1462 DEBUG_REQ(D_INFO, req, "Can't unpack body\n");
1466 /* set/clear over quota flag for a uid/gid */
1467 if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE &&
1468 body->oa.o_valid & (OBD_MD_FLUSRQUOTA | OBD_MD_FLGRPQUOTA)) {
1469 unsigned int qid[MAXQUOTAS] = { body->oa.o_uid, body->oa.o_gid };
1471 CDEBUG(D_QUOTA, "setdq for [%u %u] with valid "LPX64", flags %x\n",
1472 body->oa.o_uid, body->oa.o_gid, body->oa.o_valid,
1474 osc_quota_setdq(cli, qid, body->oa.o_valid, body->oa.o_flags);
1477 osc_update_grant(cli, body);
1482 if (aa->aa_oa->o_valid & OBD_MD_FLCKSUM)
1483 client_cksum = aa->aa_oa->o_cksum; /* save for later */
1485 if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE) {
1487 CERROR("Unexpected +ve rc %d\n", rc);
1490 LASSERT(req->rq_bulk->bd_nob == aa->aa_requested_nob);
1492 if (sptlrpc_cli_unwrap_bulk_write(req, req->rq_bulk))
1495 if ((aa->aa_oa->o_valid & OBD_MD_FLCKSUM) && client_cksum &&
1496 check_write_checksum(&body->oa, peer, client_cksum,
1497 body->oa.o_cksum, aa->aa_requested_nob,
1498 aa->aa_page_count, aa->aa_ppga,
1499 cksum_type_unpack(aa->aa_oa->o_flags)))
1502 rc = check_write_rcs(req, aa->aa_requested_nob,aa->aa_nio_count,
1503 aa->aa_page_count, aa->aa_ppga);
1507 /* The rest of this function executes only for OST_READs */
1509 /* if unwrap_bulk failed, return -EAGAIN to retry */
1510 rc = sptlrpc_cli_unwrap_bulk_read(req, req->rq_bulk, rc);
1512 GOTO(out, rc = -EAGAIN);
1514 if (rc > aa->aa_requested_nob) {
1515 CERROR("Unexpected rc %d (%d requested)\n", rc,
1516 aa->aa_requested_nob);
1520 if (rc != req->rq_bulk->bd_nob_transferred) {
1521 CERROR ("Unexpected rc %d (%d transferred)\n",
1522 rc, req->rq_bulk->bd_nob_transferred);
1526 if (rc < aa->aa_requested_nob)
1527 handle_short_read(rc, aa->aa_page_count, aa->aa_ppga);
1529 if (body->oa.o_valid & OBD_MD_FLCKSUM) {
1530 static int cksum_counter;
1531 __u32 server_cksum = body->oa.o_cksum;
1534 cksum_type_t cksum_type;
1536 cksum_type = cksum_type_unpack(body->oa.o_valid &OBD_MD_FLFLAGS?
1537 body->oa.o_flags : 0);
1538 client_cksum = osc_checksum_bulk(rc, aa->aa_page_count,
1539 aa->aa_ppga, OST_READ,
1542 if (peer->nid == req->rq_bulk->bd_sender) {
1546 router = libcfs_nid2str(req->rq_bulk->bd_sender);
1549 if (server_cksum == ~0 && rc > 0) {
1550 CERROR("Protocol error: server %s set the 'checksum' "
1551 "bit, but didn't send a checksum. Not fatal, "
1552 "but please notify on http://bugs.whamcloud.com/\n",
1553 libcfs_nid2str(peer->nid));
1554 } else if (server_cksum != client_cksum) {
1555 LCONSOLE_ERROR_MSG(0x133, "%s: BAD READ CHECKSUM: from "
1556 "%s%s%s inode "DFID" object "
1557 LPU64"/"LPU64" extent "
1558 "["LPU64"-"LPU64"]\n",
1559 req->rq_import->imp_obd->obd_name,
1560 libcfs_nid2str(peer->nid),
1562 body->oa.o_valid & OBD_MD_FLFID ?
1563 body->oa.o_parent_seq : (__u64)0,
1564 body->oa.o_valid & OBD_MD_FLFID ?
1565 body->oa.o_parent_oid : 0,
1566 body->oa.o_valid & OBD_MD_FLFID ?
1567 body->oa.o_parent_ver : 0,
1569 body->oa.o_valid & OBD_MD_FLGROUP ?
1570 body->oa.o_seq : (__u64)0,
1571 aa->aa_ppga[0]->off,
1572 aa->aa_ppga[aa->aa_page_count-1]->off +
1573 aa->aa_ppga[aa->aa_page_count-1]->count -
1575 CERROR("client %x, server %x, cksum_type %x\n",
1576 client_cksum, server_cksum, cksum_type);
1578 aa->aa_oa->o_cksum = client_cksum;
1582 CDEBUG(D_PAGE, "checksum %x confirmed\n", client_cksum);
1585 } else if (unlikely(client_cksum)) {
1586 static int cksum_missed;
1589 if ((cksum_missed & (-cksum_missed)) == cksum_missed)
1590 CERROR("Checksum %u requested from %s but not sent\n",
1591 cksum_missed, libcfs_nid2str(peer->nid));
1597 lustre_get_wire_obdo(aa->aa_oa, &body->oa);
1602 static int osc_brw_internal(int cmd, struct obd_export *exp, struct obdo *oa,
1603 struct lov_stripe_md *lsm,
1604 obd_count page_count, struct brw_page **pga,
1605 struct obd_capa *ocapa)
1607 struct ptlrpc_request *req;
1610 int generation, resends = 0;
1611 struct l_wait_info lwi;
1615 cfs_waitq_init(&waitq);
1616 generation = exp->exp_obd->u.cli.cl_import->imp_generation;
1619 rc = osc_brw_prep_request(cmd, &exp->exp_obd->u.cli, oa, lsm,
1620 page_count, pga, &req, ocapa, 0, resends);
1625 req->rq_generation_set = 1;
1626 req->rq_import_generation = generation;
1627 req->rq_sent = cfs_time_current_sec() + resends;
1630 rc = ptlrpc_queue_wait(req);
1632 if (rc == -ETIMEDOUT && req->rq_resend) {
1633 DEBUG_REQ(D_HA, req, "BULK TIMEOUT");
1634 ptlrpc_req_finished(req);
1638 rc = osc_brw_fini_request(req, rc);
1640 ptlrpc_req_finished(req);
1641 /* When server return -EINPROGRESS, client should always retry
1642 * regardless of the number of times the bulk was resent already.*/
1643 if (osc_recoverable_error(rc)) {
1645 if (rc != -EINPROGRESS &&
1646 !client_should_resend(resends, &exp->exp_obd->u.cli)) {
1647 CERROR("%s: too many resend retries for object: "
1648 ""LPU64":"LPU64", rc = %d.\n",
1649 exp->exp_obd->obd_name, oa->o_id, oa->o_seq, rc);
1653 exp->exp_obd->u.cli.cl_import->imp_generation) {
1654 CDEBUG(D_HA, "%s: resend cross eviction for object: "
1655 ""LPU64":"LPU64", rc = %d.\n",
1656 exp->exp_obd->obd_name, oa->o_id, oa->o_seq, rc);
1660 lwi = LWI_TIMEOUT_INTR(cfs_time_seconds(resends), NULL, NULL,
1662 l_wait_event(waitq, 0, &lwi);
1667 if (rc == -EAGAIN || rc == -EINPROGRESS)
1672 int osc_brw_redo_request(struct ptlrpc_request *request,
1673 struct osc_brw_async_args *aa)
1675 struct ptlrpc_request *new_req;
1676 struct ptlrpc_request_set *set = request->rq_set;
1677 struct osc_brw_async_args *new_aa;
1678 struct osc_async_page *oap;
1682 DEBUG_REQ(D_ERROR, request, "redo for recoverable error");
1684 rc = osc_brw_prep_request(lustre_msg_get_opc(request->rq_reqmsg) ==
1685 OST_WRITE ? OBD_BRW_WRITE :OBD_BRW_READ,
1686 aa->aa_cli, aa->aa_oa,
1687 NULL /* lsm unused by osc currently */,
1688 aa->aa_page_count, aa->aa_ppga,
1689 &new_req, aa->aa_ocapa, 0, 1);
1693 client_obd_list_lock(&aa->aa_cli->cl_loi_list_lock);
1695 cfs_list_for_each_entry(oap, &aa->aa_oaps, oap_rpc_item) {
1696 if (oap->oap_request != NULL) {
1697 LASSERTF(request == oap->oap_request,
1698 "request %p != oap_request %p\n",
1699 request, oap->oap_request);
1700 if (oap->oap_interrupted) {
1701 client_obd_list_unlock(&aa->aa_cli->cl_loi_list_lock);
1702 ptlrpc_req_finished(new_req);
1707 /* New request takes over pga and oaps from old request.
1708 * Note that copying a list_head doesn't work, need to move it... */
1710 new_req->rq_interpret_reply = request->rq_interpret_reply;
1711 new_req->rq_async_args = request->rq_async_args;
1712 new_req->rq_sent = cfs_time_current_sec() + aa->aa_resends;
1713 new_req->rq_generation_set = 1;
1714 new_req->rq_import_generation = request->rq_import_generation;
1716 new_aa = ptlrpc_req_async_args(new_req);
1718 CFS_INIT_LIST_HEAD(&new_aa->aa_oaps);
1719 cfs_list_splice(&aa->aa_oaps, &new_aa->aa_oaps);
1720 CFS_INIT_LIST_HEAD(&aa->aa_oaps);
1722 cfs_list_for_each_entry(oap, &new_aa->aa_oaps, oap_rpc_item) {
1723 if (oap->oap_request) {
1724 ptlrpc_req_finished(oap->oap_request);
1725 oap->oap_request = ptlrpc_request_addref(new_req);
1729 new_aa->aa_ocapa = aa->aa_ocapa;
1730 aa->aa_ocapa = NULL;
1732 /* use ptlrpc_set_add_req is safe because interpret functions work
1733 * in check_set context. only one way exist with access to request
1734 * from different thread got -EINTR - this way protected with
1735 * cl_loi_list_lock */
1736 ptlrpc_set_add_req(set, new_req);
1738 client_obd_list_unlock(&aa->aa_cli->cl_loi_list_lock);
1740 DEBUG_REQ(D_INFO, new_req, "new request");
1745 * ugh, we want disk allocation on the target to happen in offset order. we'll
1746 * follow sedgewicks advice and stick to the dead simple shellsort -- it'll do
1747 * fine for our small page arrays and doesn't require allocation. its an
1748 * insertion sort that swaps elements that are strides apart, shrinking the
1749 * stride down until its '1' and the array is sorted.
1751 static void sort_brw_pages(struct brw_page **array, int num)
1754 struct brw_page *tmp;
1758 for (stride = 1; stride < num ; stride = (stride * 3) + 1)
1763 for (i = stride ; i < num ; i++) {
1766 while (j >= stride && array[j - stride]->off > tmp->off) {
1767 array[j] = array[j - stride];
1772 } while (stride > 1);
1775 static obd_count max_unfragmented_pages(struct brw_page **pg, obd_count pages)
1781 LASSERT (pages > 0);
1782 offset = pg[i]->off & ~CFS_PAGE_MASK;
1786 if (pages == 0) /* that's all */
1789 if (offset + pg[i]->count < CFS_PAGE_SIZE)
1790 return count; /* doesn't end on page boundary */
1793 offset = pg[i]->off & ~CFS_PAGE_MASK;
1794 if (offset != 0) /* doesn't start on page boundary */
1801 static struct brw_page **osc_build_ppga(struct brw_page *pga, obd_count count)
1803 struct brw_page **ppga;
1806 OBD_ALLOC(ppga, sizeof(*ppga) * count);
1810 for (i = 0; i < count; i++)
1815 static void osc_release_ppga(struct brw_page **ppga, obd_count count)
1817 LASSERT(ppga != NULL);
1818 OBD_FREE(ppga, sizeof(*ppga) * count);
1821 static int osc_brw(int cmd, struct obd_export *exp, struct obd_info *oinfo,
1822 obd_count page_count, struct brw_page *pga,
1823 struct obd_trans_info *oti)
1825 struct obdo *saved_oa = NULL;
1826 struct brw_page **ppga, **orig;
1827 struct obd_import *imp = class_exp2cliimp(exp);
1828 struct client_obd *cli;
1829 int rc, page_count_orig;
1832 LASSERT((imp != NULL) && (imp->imp_obd != NULL));
1833 cli = &imp->imp_obd->u.cli;
1835 if (cmd & OBD_BRW_CHECK) {
1836 /* The caller just wants to know if there's a chance that this
1837 * I/O can succeed */
1839 if (imp->imp_invalid)
1844 /* test_brw with a failed create can trip this, maybe others. */
1845 LASSERT(cli->cl_max_pages_per_rpc);
1849 orig = ppga = osc_build_ppga(pga, page_count);
1852 page_count_orig = page_count;
1854 sort_brw_pages(ppga, page_count);
1855 while (page_count) {
1856 obd_count pages_per_brw;
1858 if (page_count > cli->cl_max_pages_per_rpc)
1859 pages_per_brw = cli->cl_max_pages_per_rpc;
1861 pages_per_brw = page_count;
1863 pages_per_brw = max_unfragmented_pages(ppga, pages_per_brw);
1865 if (saved_oa != NULL) {
1866 /* restore previously saved oa */
1867 *oinfo->oi_oa = *saved_oa;
1868 } else if (page_count > pages_per_brw) {
1869 /* save a copy of oa (brw will clobber it) */
1870 OBDO_ALLOC(saved_oa);
1871 if (saved_oa == NULL)
1872 GOTO(out, rc = -ENOMEM);
1873 *saved_oa = *oinfo->oi_oa;
1876 rc = osc_brw_internal(cmd, exp, oinfo->oi_oa, oinfo->oi_md,
1877 pages_per_brw, ppga, oinfo->oi_capa);
1882 page_count -= pages_per_brw;
1883 ppga += pages_per_brw;
1887 osc_release_ppga(orig, page_count_orig);
1889 if (saved_oa != NULL)
1890 OBDO_FREE(saved_oa);
1895 static int brw_interpret(const struct lu_env *env,
1896 struct ptlrpc_request *req, void *data, int rc)
1898 struct osc_brw_async_args *aa = data;
1899 struct osc_async_page *oap, *tmp;
1900 struct client_obd *cli;
1903 rc = osc_brw_fini_request(req, rc);
1904 CDEBUG(D_INODE, "request %p aa %p rc %d\n", req, aa, rc);
1905 /* When server return -EINPROGRESS, client should always retry
1906 * regardless of the number of times the bulk was resent already. */
1907 if (osc_recoverable_error(rc)) {
1908 if (req->rq_import_generation !=
1909 req->rq_import->imp_generation) {
1910 CDEBUG(D_HA, "%s: resend cross eviction for object: "
1911 ""LPU64":"LPU64", rc = %d.\n",
1912 req->rq_import->imp_obd->obd_name,
1913 aa->aa_oa->o_id, aa->aa_oa->o_seq, rc);
1914 } else if (rc == -EINPROGRESS ||
1915 client_should_resend(aa->aa_resends, aa->aa_cli)) {
1916 rc = osc_brw_redo_request(req, aa);
1918 CERROR("%s: too many resent retries for object: "
1919 ""LPU64":"LPU64", rc = %d.\n",
1920 req->rq_import->imp_obd->obd_name,
1921 aa->aa_oa->o_id, aa->aa_oa->o_seq, rc);
1926 else if (rc == -EAGAIN || rc == -EINPROGRESS)
1931 capa_put(aa->aa_ocapa);
1932 aa->aa_ocapa = NULL;
1936 client_obd_list_lock(&cli->cl_loi_list_lock);
1938 /* We need to decrement before osc_ap_completion->osc_wake_cache_waiters
1939 * is called so we know whether to go to sync BRWs or wait for more
1940 * RPCs to complete */
1941 if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE)
1942 cli->cl_w_in_flight--;
1944 cli->cl_r_in_flight--;
1946 /* the caller may re-use the oap after the completion call so
1947 * we need to clean it up a little */
1948 cfs_list_for_each_entry_safe(oap, tmp, &aa->aa_oaps,
1950 cfs_list_del_init(&oap->oap_rpc_item);
1951 osc_ap_completion(env, cli, aa->aa_oa, oap, 1, rc);
1953 OBDO_FREE(aa->aa_oa);
1955 osc_wake_cache_waiters(cli);
1956 osc_io_unplug(env, cli, NULL, PDL_POLICY_SAME);
1957 client_obd_list_unlock(&cli->cl_loi_list_lock);
1959 cl_req_completion(env, aa->aa_clerq, rc < 0 ? rc :
1960 req->rq_bulk->bd_nob_transferred);
1961 osc_release_ppga(aa->aa_ppga, aa->aa_page_count);
1962 ptlrpc_lprocfs_brw(req, req->rq_bulk->bd_nob_transferred);
1967 /* The most tricky part of this function is that it will return with
1968 * cli->cli_loi_list_lock held.
1970 int osc_build_rpc(const struct lu_env *env, struct client_obd *cli,
1971 cfs_list_t *rpc_list, int page_count, int cmd,
1974 struct ptlrpc_request *req = NULL;
1975 struct brw_page **pga = NULL;
1976 struct osc_brw_async_args *aa = NULL;
1977 struct obdo *oa = NULL;
1978 struct osc_async_page *oap;
1979 struct osc_async_page *tmp;
1980 struct cl_req *clerq = NULL;
1981 enum cl_req_type crt = (cmd & OBD_BRW_WRITE) ? CRT_WRITE : CRT_READ;
1982 struct ldlm_lock *lock = NULL;
1983 struct cl_req_attr crattr;
1984 int i, rc, mpflag = 0;
1987 LASSERT(!cfs_list_empty(rpc_list));
1989 if (cmd & OBD_BRW_MEMALLOC)
1990 mpflag = cfs_memory_pressure_get_and_set();
1992 memset(&crattr, 0, sizeof crattr);
1993 OBD_ALLOC(pga, sizeof(*pga) * page_count);
1995 GOTO(out, rc = -ENOMEM);
1999 GOTO(out, rc = -ENOMEM);
2002 cfs_list_for_each_entry(oap, rpc_list, oap_rpc_item) {
2003 struct cl_page *page = osc_oap2cl_page(oap);
2004 if (clerq == NULL) {
2005 clerq = cl_req_alloc(env, page, crt,
2006 1 /* only 1-object rpcs for
2009 GOTO(out, rc = PTR_ERR(clerq));
2010 lock = oap->oap_ldlm_lock;
2012 pga[i] = &oap->oap_brw_page;
2013 pga[i]->off = oap->oap_obj_off + oap->oap_page_off;
2014 CDEBUG(0, "put page %p index %lu oap %p flg %x to pga\n",
2015 pga[i]->pg, cfs_page_index(oap->oap_page), oap, pga[i]->flag);
2017 cl_req_page_add(env, clerq, page);
2020 /* always get the data for the obdo for the rpc */
2021 LASSERT(clerq != NULL);
2023 crattr.cra_capa = NULL;
2024 memset(crattr.cra_jobid, 0, JOBSTATS_JOBID_SIZE);
2025 cl_req_attr_set(env, clerq, &crattr, ~0ULL);
2027 oa->o_handle = lock->l_remote_handle;
2028 oa->o_valid |= OBD_MD_FLHANDLE;
2031 rc = cl_req_prep(env, clerq);
2033 CERROR("cl_req_prep failed: %d\n", rc);
2037 sort_brw_pages(pga, page_count);
2038 rc = osc_brw_prep_request(cmd, cli, oa, NULL, page_count,
2039 pga, &req, crattr.cra_capa, 1, 0);
2041 CERROR("prep_req failed: %d\n", rc);
2045 req->rq_interpret_reply = brw_interpret;
2046 if (cmd & OBD_BRW_MEMALLOC)
2047 req->rq_memalloc = 1;
2049 /* Need to update the timestamps after the request is built in case
2050 * we race with setattr (locally or in queue at OST). If OST gets
2051 * later setattr before earlier BRW (as determined by the request xid),
2052 * the OST will not use BRW timestamps. Sadly, there is no obvious
2053 * way to do this in a single call. bug 10150 */
2054 cl_req_attr_set(env, clerq, &crattr,
2055 OBD_MD_FLMTIME|OBD_MD_FLCTIME|OBD_MD_FLATIME);
2057 lustre_msg_set_jobid(req->rq_reqmsg, crattr.cra_jobid);
2059 CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
2060 aa = ptlrpc_req_async_args(req);
2061 CFS_INIT_LIST_HEAD(&aa->aa_oaps);
2062 cfs_list_splice(rpc_list, &aa->aa_oaps);
2063 CFS_INIT_LIST_HEAD(rpc_list);
2064 aa->aa_clerq = clerq;
2066 if (cmd & OBD_BRW_MEMALLOC)
2067 cfs_memory_pressure_restore(mpflag);
2069 capa_put(crattr.cra_capa);
2071 LASSERT(req == NULL);
2076 OBD_FREE(pga, sizeof(*pga) * page_count);
2077 /* this should happen rarely and is pretty bad, it makes the
2078 * pending list not follow the dirty order */
2079 client_obd_list_lock(&cli->cl_loi_list_lock);
2080 cfs_list_for_each_entry_safe(oap, tmp, rpc_list, oap_rpc_item) {
2081 cfs_list_del_init(&oap->oap_rpc_item);
2083 /* queued sync pages can be torn down while the pages
2084 * were between the pending list and the rpc */
2085 if (oap->oap_interrupted) {
2086 CDEBUG(D_INODE, "oap %p interrupted\n", oap);
2087 osc_ap_completion(env, cli, NULL, oap, 0,
2091 osc_ap_completion(env, cli, NULL, oap, 0, rc);
2093 if (clerq && !IS_ERR(clerq))
2094 cl_req_completion(env, clerq, rc);
2096 struct osc_async_page *tmp = NULL;
2098 /* queued sync pages can be torn down while the pages
2099 * were between the pending list and the rpc */
2100 LASSERT(aa != NULL);
2101 client_obd_list_lock(&cli->cl_loi_list_lock);
2102 cfs_list_for_each_entry(oap, &aa->aa_oaps, oap_rpc_item) {
2103 /* only one oap gets a request reference */
2106 if (oap->oap_interrupted && !req->rq_intr) {
2107 CDEBUG(D_INODE, "oap %p in req %p interrupted\n",
2109 ptlrpc_mark_interrupted(req);
2113 tmp->oap_request = ptlrpc_request_addref(req);
2115 DEBUG_REQ(D_INODE,req, "%d pages, aa %p. now %dr/%dw in flight",
2116 page_count, aa, cli->cl_r_in_flight,
2117 cli->cl_w_in_flight);
2119 /* XXX: Maybe the caller can check the RPC bulk descriptor to
2120 * see which CPU/NUMA node the majority of pages were allocated
2121 * on, and try to assign the async RPC to the CPU core
2122 * (PDL_POLICY_PREFERRED) to reduce cross-CPU memory traffic.
2124 * But on the other hand, we expect that multiple ptlrpcd
2125 * threads and the initial write sponsor can run in parallel,
2126 * especially when data checksum is enabled, which is CPU-bound
2127 * operation and single ptlrpcd thread cannot process in time.
2128 * So more ptlrpcd threads sharing BRW load
2129 * (with PDL_POLICY_ROUND) seems better.
2131 ptlrpcd_add_req(req, pol, -1);
2136 static int osc_set_lock_data_with_check(struct ldlm_lock *lock,
2137 struct ldlm_enqueue_info *einfo)
2139 void *data = einfo->ei_cbdata;
2142 LASSERT(lock != NULL);
2143 LASSERT(lock->l_blocking_ast == einfo->ei_cb_bl);
2144 LASSERT(lock->l_resource->lr_type == einfo->ei_type);
2145 LASSERT(lock->l_completion_ast == einfo->ei_cb_cp);
2146 LASSERT(lock->l_glimpse_ast == einfo->ei_cb_gl);
2148 lock_res_and_lock(lock);
2149 cfs_spin_lock(&osc_ast_guard);
2151 if (lock->l_ast_data == NULL)
2152 lock->l_ast_data = data;
2153 if (lock->l_ast_data == data)
2156 cfs_spin_unlock(&osc_ast_guard);
2157 unlock_res_and_lock(lock);
2162 static int osc_set_data_with_check(struct lustre_handle *lockh,
2163 struct ldlm_enqueue_info *einfo)
2165 struct ldlm_lock *lock = ldlm_handle2lock(lockh);
2169 set = osc_set_lock_data_with_check(lock, einfo);
2170 LDLM_LOCK_PUT(lock);
2172 CERROR("lockh %p, data %p - client evicted?\n",
2173 lockh, einfo->ei_cbdata);
2177 static int osc_change_cbdata(struct obd_export *exp, struct lov_stripe_md *lsm,
2178 ldlm_iterator_t replace, void *data)
2180 struct ldlm_res_id res_id;
2181 struct obd_device *obd = class_exp2obd(exp);
2183 osc_build_res_name(lsm->lsm_object_id, lsm->lsm_object_seq, &res_id);
2184 ldlm_resource_iterate(obd->obd_namespace, &res_id, replace, data);
2188 /* find any ldlm lock of the inode in osc
2192 static int osc_find_cbdata(struct obd_export *exp, struct lov_stripe_md *lsm,
2193 ldlm_iterator_t replace, void *data)
2195 struct ldlm_res_id res_id;
2196 struct obd_device *obd = class_exp2obd(exp);
2199 osc_build_res_name(lsm->lsm_object_id, lsm->lsm_object_seq, &res_id);
2200 rc = ldlm_resource_iterate(obd->obd_namespace, &res_id, replace, data);
2201 if (rc == LDLM_ITER_STOP)
2203 if (rc == LDLM_ITER_CONTINUE)
2208 static int osc_enqueue_fini(struct ptlrpc_request *req, struct ost_lvb *lvb,
2209 obd_enqueue_update_f upcall, void *cookie,
2210 int *flags, int agl, int rc)
2212 int intent = *flags & LDLM_FL_HAS_INTENT;
2216 /* The request was created before ldlm_cli_enqueue call. */
2217 if (rc == ELDLM_LOCK_ABORTED) {
2218 struct ldlm_reply *rep;
2219 rep = req_capsule_server_get(&req->rq_pill,
2222 LASSERT(rep != NULL);
2223 if (rep->lock_policy_res1)
2224 rc = rep->lock_policy_res1;
2228 if ((intent != 0 && rc == ELDLM_LOCK_ABORTED && agl == 0) ||
2230 *flags |= LDLM_FL_LVB_READY;
2231 CDEBUG(D_INODE,"got kms "LPU64" blocks "LPU64" mtime "LPU64"\n",
2232 lvb->lvb_size, lvb->lvb_blocks, lvb->lvb_mtime);
2235 /* Call the update callback. */
2236 rc = (*upcall)(cookie, rc);
2240 static int osc_enqueue_interpret(const struct lu_env *env,
2241 struct ptlrpc_request *req,
2242 struct osc_enqueue_args *aa, int rc)
2244 struct ldlm_lock *lock;
2245 struct lustre_handle handle;
2247 struct ost_lvb *lvb;
2249 int *flags = aa->oa_flags;
2251 /* Make a local copy of a lock handle and a mode, because aa->oa_*
2252 * might be freed anytime after lock upcall has been called. */
2253 lustre_handle_copy(&handle, aa->oa_lockh);
2254 mode = aa->oa_ei->ei_mode;
2256 /* ldlm_cli_enqueue is holding a reference on the lock, so it must
2258 lock = ldlm_handle2lock(&handle);
2260 /* Take an additional reference so that a blocking AST that
2261 * ldlm_cli_enqueue_fini() might post for a failed lock, is guaranteed
2262 * to arrive after an upcall has been executed by
2263 * osc_enqueue_fini(). */
2264 ldlm_lock_addref(&handle, mode);
2266 /* Let CP AST to grant the lock first. */
2267 OBD_FAIL_TIMEOUT(OBD_FAIL_OSC_CP_ENQ_RACE, 1);
2269 if (aa->oa_agl && rc == ELDLM_LOCK_ABORTED) {
2274 lvb_len = sizeof(*aa->oa_lvb);
2277 /* Complete obtaining the lock procedure. */
2278 rc = ldlm_cli_enqueue_fini(aa->oa_exp, req, aa->oa_ei->ei_type, 1,
2279 mode, flags, lvb, lvb_len, &handle, rc);
2280 /* Complete osc stuff. */
2281 rc = osc_enqueue_fini(req, aa->oa_lvb, aa->oa_upcall, aa->oa_cookie,
2282 flags, aa->oa_agl, rc);
2284 OBD_FAIL_TIMEOUT(OBD_FAIL_OSC_CP_CANCEL_RACE, 10);
2286 /* Release the lock for async request. */
2287 if (lustre_handle_is_used(&handle) && rc == ELDLM_OK)
2289 * Releases a reference taken by ldlm_cli_enqueue(), if it is
2290 * not already released by
2291 * ldlm_cli_enqueue_fini()->failed_lock_cleanup()
2293 ldlm_lock_decref(&handle, mode);
2295 LASSERTF(lock != NULL, "lockh %p, req %p, aa %p - client evicted?\n",
2296 aa->oa_lockh, req, aa);
2297 ldlm_lock_decref(&handle, mode);
2298 LDLM_LOCK_PUT(lock);
2302 void osc_update_enqueue(struct lustre_handle *lov_lockhp,
2303 struct lov_oinfo *loi, int flags,
2304 struct ost_lvb *lvb, __u32 mode, int rc)
2306 struct ldlm_lock *lock = ldlm_handle2lock(lov_lockhp);
2308 if (rc == ELDLM_OK) {
2311 LASSERT(lock != NULL);
2312 loi->loi_lvb = *lvb;
2313 tmp = loi->loi_lvb.lvb_size;
2314 /* Extend KMS up to the end of this lock and no further
2315 * A lock on [x,y] means a KMS of up to y + 1 bytes! */
2316 if (tmp > lock->l_policy_data.l_extent.end)
2317 tmp = lock->l_policy_data.l_extent.end + 1;
2318 if (tmp >= loi->loi_kms) {
2319 LDLM_DEBUG(lock, "lock acquired, setting rss="LPU64
2320 ", kms="LPU64, loi->loi_lvb.lvb_size, tmp);
2321 loi_kms_set(loi, tmp);
2323 LDLM_DEBUG(lock, "lock acquired, setting rss="
2324 LPU64"; leaving kms="LPU64", end="LPU64,
2325 loi->loi_lvb.lvb_size, loi->loi_kms,
2326 lock->l_policy_data.l_extent.end);
2328 ldlm_lock_allow_match(lock);
2329 } else if (rc == ELDLM_LOCK_ABORTED && (flags & LDLM_FL_HAS_INTENT)) {
2330 LASSERT(lock != NULL);
2331 loi->loi_lvb = *lvb;
2332 ldlm_lock_allow_match(lock);
2333 CDEBUG(D_INODE, "glimpsed, setting rss="LPU64"; leaving"
2334 " kms="LPU64"\n", loi->loi_lvb.lvb_size, loi->loi_kms);
2340 ldlm_lock_fail_match(lock);
2342 LDLM_LOCK_PUT(lock);
2345 EXPORT_SYMBOL(osc_update_enqueue);
2347 struct ptlrpc_request_set *PTLRPCD_SET = (void *)1;
2349 /* When enqueuing asynchronously, locks are not ordered, we can obtain a lock
2350 * from the 2nd OSC before a lock from the 1st one. This does not deadlock with
2351 * other synchronous requests, however keeping some locks and trying to obtain
2352 * others may take a considerable amount of time in a case of ost failure; and
2353 * when other sync requests do not get released lock from a client, the client
2354 * is excluded from the cluster -- such scenarious make the life difficult, so
2355 * release locks just after they are obtained. */
2356 int osc_enqueue_base(struct obd_export *exp, struct ldlm_res_id *res_id,
2357 int *flags, ldlm_policy_data_t *policy,
2358 struct ost_lvb *lvb, int kms_valid,
2359 obd_enqueue_update_f upcall, void *cookie,
2360 struct ldlm_enqueue_info *einfo,
2361 struct lustre_handle *lockh,
2362 struct ptlrpc_request_set *rqset, int async, int agl)
2364 struct obd_device *obd = exp->exp_obd;
2365 struct ptlrpc_request *req = NULL;
2366 int intent = *flags & LDLM_FL_HAS_INTENT;
2367 int match_lvb = (agl != 0 ? 0 : LDLM_FL_LVB_READY);
2372 /* Filesystem lock extents are extended to page boundaries so that
2373 * dealing with the page cache is a little smoother. */
2374 policy->l_extent.start -= policy->l_extent.start & ~CFS_PAGE_MASK;
2375 policy->l_extent.end |= ~CFS_PAGE_MASK;
2378 * kms is not valid when either object is completely fresh (so that no
2379 * locks are cached), or object was evicted. In the latter case cached
2380 * lock cannot be used, because it would prime inode state with
2381 * potentially stale LVB.
2386 /* Next, search for already existing extent locks that will cover us */
2387 /* If we're trying to read, we also search for an existing PW lock. The
2388 * VFS and page cache already protect us locally, so lots of readers/
2389 * writers can share a single PW lock.
2391 * There are problems with conversion deadlocks, so instead of
2392 * converting a read lock to a write lock, we'll just enqueue a new
2395 * At some point we should cancel the read lock instead of making them
2396 * send us a blocking callback, but there are problems with canceling
2397 * locks out from other users right now, too. */
2398 mode = einfo->ei_mode;
2399 if (einfo->ei_mode == LCK_PR)
2401 mode = ldlm_lock_match(obd->obd_namespace, *flags | match_lvb, res_id,
2402 einfo->ei_type, policy, mode, lockh, 0);
2404 struct ldlm_lock *matched = ldlm_handle2lock(lockh);
2406 if ((agl != 0) && !(matched->l_flags & LDLM_FL_LVB_READY)) {
2407 /* For AGL, if enqueue RPC is sent but the lock is not
2408 * granted, then skip to process this strpe.
2409 * Return -ECANCELED to tell the caller. */
2410 ldlm_lock_decref(lockh, mode);
2411 LDLM_LOCK_PUT(matched);
2413 } else if (osc_set_lock_data_with_check(matched, einfo)) {
2414 *flags |= LDLM_FL_LVB_READY;
2415 /* addref the lock only if not async requests and PW
2416 * lock is matched whereas we asked for PR. */
2417 if (!rqset && einfo->ei_mode != mode)
2418 ldlm_lock_addref(lockh, LCK_PR);
2420 /* I would like to be able to ASSERT here that
2421 * rss <= kms, but I can't, for reasons which
2422 * are explained in lov_enqueue() */
2425 /* We already have a lock, and it's referenced */
2426 (*upcall)(cookie, ELDLM_OK);
2428 if (einfo->ei_mode != mode)
2429 ldlm_lock_decref(lockh, LCK_PW);
2431 /* For async requests, decref the lock. */
2432 ldlm_lock_decref(lockh, einfo->ei_mode);
2433 LDLM_LOCK_PUT(matched);
2436 ldlm_lock_decref(lockh, mode);
2437 LDLM_LOCK_PUT(matched);
2443 CFS_LIST_HEAD(cancels);
2444 req = ptlrpc_request_alloc(class_exp2cliimp(exp),
2445 &RQF_LDLM_ENQUEUE_LVB);
2449 rc = ldlm_prep_enqueue_req(exp, req, &cancels, 0);
2451 ptlrpc_request_free(req);
2455 req_capsule_set_size(&req->rq_pill, &RMF_DLM_LVB, RCL_SERVER,
2457 ptlrpc_request_set_replen(req);
2460 /* users of osc_enqueue() can pass this flag for ldlm_lock_match() */
2461 *flags &= ~LDLM_FL_BLOCK_GRANTED;
2463 rc = ldlm_cli_enqueue(exp, &req, einfo, res_id, policy, flags, lvb,
2464 sizeof(*lvb), lockh, async);
2467 struct osc_enqueue_args *aa;
2468 CLASSERT (sizeof(*aa) <= sizeof(req->rq_async_args));
2469 aa = ptlrpc_req_async_args(req);
2472 aa->oa_flags = flags;
2473 aa->oa_upcall = upcall;
2474 aa->oa_cookie = cookie;
2476 aa->oa_lockh = lockh;
2479 req->rq_interpret_reply =
2480 (ptlrpc_interpterer_t)osc_enqueue_interpret;
2481 if (rqset == PTLRPCD_SET)
2482 ptlrpcd_add_req(req, PDL_POLICY_ROUND, -1);
2484 ptlrpc_set_add_req(rqset, req);
2485 } else if (intent) {
2486 ptlrpc_req_finished(req);
2491 rc = osc_enqueue_fini(req, lvb, upcall, cookie, flags, agl, rc);
2493 ptlrpc_req_finished(req);
2498 static int osc_enqueue(struct obd_export *exp, struct obd_info *oinfo,
2499 struct ldlm_enqueue_info *einfo,
2500 struct ptlrpc_request_set *rqset)
2502 struct ldlm_res_id res_id;
2506 osc_build_res_name(oinfo->oi_md->lsm_object_id,
2507 oinfo->oi_md->lsm_object_seq, &res_id);
2509 rc = osc_enqueue_base(exp, &res_id, &oinfo->oi_flags, &oinfo->oi_policy,
2510 &oinfo->oi_md->lsm_oinfo[0]->loi_lvb,
2511 oinfo->oi_md->lsm_oinfo[0]->loi_kms_valid,
2512 oinfo->oi_cb_up, oinfo, einfo, oinfo->oi_lockh,
2513 rqset, rqset != NULL, 0);
2517 int osc_match_base(struct obd_export *exp, struct ldlm_res_id *res_id,
2518 __u32 type, ldlm_policy_data_t *policy, __u32 mode,
2519 int *flags, void *data, struct lustre_handle *lockh,
2522 struct obd_device *obd = exp->exp_obd;
2523 int lflags = *flags;
2527 if (OBD_FAIL_CHECK(OBD_FAIL_OSC_MATCH))
2530 /* Filesystem lock extents are extended to page boundaries so that
2531 * dealing with the page cache is a little smoother */
2532 policy->l_extent.start -= policy->l_extent.start & ~CFS_PAGE_MASK;
2533 policy->l_extent.end |= ~CFS_PAGE_MASK;
2535 /* Next, search for already existing extent locks that will cover us */
2536 /* If we're trying to read, we also search for an existing PW lock. The
2537 * VFS and page cache already protect us locally, so lots of readers/
2538 * writers can share a single PW lock. */
2542 rc = ldlm_lock_match(obd->obd_namespace, lflags,
2543 res_id, type, policy, rc, lockh, unref);
2546 if (!osc_set_data_with_check(lockh, data)) {
2547 if (!(lflags & LDLM_FL_TEST_LOCK))
2548 ldlm_lock_decref(lockh, rc);
2552 if (!(lflags & LDLM_FL_TEST_LOCK) && mode != rc) {
2553 ldlm_lock_addref(lockh, LCK_PR);
2554 ldlm_lock_decref(lockh, LCK_PW);
2561 int osc_cancel_base(struct lustre_handle *lockh, __u32 mode)
2565 if (unlikely(mode == LCK_GROUP))
2566 ldlm_lock_decref_and_cancel(lockh, mode);
2568 ldlm_lock_decref(lockh, mode);
2573 static int osc_cancel(struct obd_export *exp, struct lov_stripe_md *md,
2574 __u32 mode, struct lustre_handle *lockh)
2577 RETURN(osc_cancel_base(lockh, mode));
2580 static int osc_cancel_unused(struct obd_export *exp,
2581 struct lov_stripe_md *lsm,
2582 ldlm_cancel_flags_t flags,
2585 struct obd_device *obd = class_exp2obd(exp);
2586 struct ldlm_res_id res_id, *resp = NULL;
2589 resp = osc_build_res_name(lsm->lsm_object_id,
2590 lsm->lsm_object_seq, &res_id);
2593 return ldlm_cli_cancel_unused(obd->obd_namespace, resp, flags, opaque);
2596 static int osc_statfs_interpret(const struct lu_env *env,
2597 struct ptlrpc_request *req,
2598 struct osc_async_args *aa, int rc)
2600 struct client_obd *cli = &req->rq_import->imp_obd->u.cli;
2601 struct obd_statfs *msfs;
2606 /* The request has in fact never been sent
2607 * due to issues at a higher level (LOV).
2608 * Exit immediately since the caller is
2609 * aware of the problem and takes care
2610 * of the clean up */
2613 if ((rc == -ENOTCONN || rc == -EAGAIN) &&
2614 (aa->aa_oi->oi_flags & OBD_STATFS_NODELAY))
2620 msfs = req_capsule_server_get(&req->rq_pill, &RMF_OBD_STATFS);
2622 GOTO(out, rc = -EPROTO);
2625 /* Reinitialize the RDONLY and DEGRADED flags at the client
2626 * on each statfs, so they don't stay set permanently. */
2627 cfs_spin_lock(&cli->cl_oscc.oscc_lock);
2629 if (unlikely(msfs->os_state & OS_STATE_DEGRADED))
2630 cli->cl_oscc.oscc_flags |= OSCC_FLAG_DEGRADED;
2631 else if (unlikely(cli->cl_oscc.oscc_flags & OSCC_FLAG_DEGRADED))
2632 cli->cl_oscc.oscc_flags &= ~OSCC_FLAG_DEGRADED;
2634 if (unlikely(msfs->os_state & OS_STATE_READONLY))
2635 cli->cl_oscc.oscc_flags |= OSCC_FLAG_RDONLY;
2636 else if (unlikely(cli->cl_oscc.oscc_flags & OSCC_FLAG_RDONLY))
2637 cli->cl_oscc.oscc_flags &= ~OSCC_FLAG_RDONLY;
2639 /* Add a bit of hysteresis so this flag isn't continually flapping,
2640 * and ensure that new files don't get extremely fragmented due to
2641 * only a small amount of available space in the filesystem.
2642 * We want to set the NOSPC flag when there is less than ~0.1% free
2643 * and clear it when there is at least ~0.2% free space, so:
2644 * avail < ~0.1% max max = avail + used
2645 * 1025 * avail < avail + used used = blocks - free
2646 * 1024 * avail < used
2647 * 1024 * avail < blocks - free
2648 * avail < ((blocks - free) >> 10)
2650 * On very large disk, say 16TB 0.1% will be 16 GB. We don't want to
2651 * lose that amount of space so in those cases we report no space left
2652 * if their is less than 1 GB left. */
2653 used = min_t(__u64,(msfs->os_blocks - msfs->os_bfree) >> 10, 1 << 30);
2654 if (unlikely(((cli->cl_oscc.oscc_flags & OSCC_FLAG_NOSPC) == 0) &&
2655 ((msfs->os_ffree < 32) || (msfs->os_bavail < used))))
2656 cli->cl_oscc.oscc_flags |= OSCC_FLAG_NOSPC;
2657 else if (unlikely(((cli->cl_oscc.oscc_flags & OSCC_FLAG_NOSPC) != 0) &&
2658 (msfs->os_ffree > 64) &&
2659 (msfs->os_bavail > (used << 1)))) {
2660 cli->cl_oscc.oscc_flags &= ~(OSCC_FLAG_NOSPC |
2661 OSCC_FLAG_NOSPC_BLK);
2664 if (unlikely(((cli->cl_oscc.oscc_flags & OSCC_FLAG_NOSPC) != 0) &&
2665 (msfs->os_bavail < used)))
2666 cli->cl_oscc.oscc_flags |= OSCC_FLAG_NOSPC_BLK;
2668 cfs_spin_unlock(&cli->cl_oscc.oscc_lock);
2670 *aa->aa_oi->oi_osfs = *msfs;
2672 rc = aa->aa_oi->oi_cb_up(aa->aa_oi, rc);
2676 static int osc_statfs_async(struct obd_export *exp,
2677 struct obd_info *oinfo, __u64 max_age,
2678 struct ptlrpc_request_set *rqset)
2680 struct obd_device *obd = class_exp2obd(exp);
2681 struct ptlrpc_request *req;
2682 struct osc_async_args *aa;
2686 /* We could possibly pass max_age in the request (as an absolute
2687 * timestamp or a "seconds.usec ago") so the target can avoid doing
2688 * extra calls into the filesystem if that isn't necessary (e.g.
2689 * during mount that would help a bit). Having relative timestamps
2690 * is not so great if request processing is slow, while absolute
2691 * timestamps are not ideal because they need time synchronization. */
2692 req = ptlrpc_request_alloc(obd->u.cli.cl_import, &RQF_OST_STATFS);
2696 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_STATFS);
2698 ptlrpc_request_free(req);
2701 ptlrpc_request_set_replen(req);
2702 req->rq_request_portal = OST_CREATE_PORTAL;
2703 ptlrpc_at_set_req_timeout(req);
2705 if (oinfo->oi_flags & OBD_STATFS_NODELAY) {
2706 /* procfs requests not want stat in wait for avoid deadlock */
2707 req->rq_no_resend = 1;
2708 req->rq_no_delay = 1;
2711 req->rq_interpret_reply = (ptlrpc_interpterer_t)osc_statfs_interpret;
2712 CLASSERT (sizeof(*aa) <= sizeof(req->rq_async_args));
2713 aa = ptlrpc_req_async_args(req);
2716 ptlrpc_set_add_req(rqset, req);
2720 static int osc_statfs(const struct lu_env *env, struct obd_export *exp,
2721 struct obd_statfs *osfs, __u64 max_age, __u32 flags)
2723 struct obd_device *obd = class_exp2obd(exp);
2724 struct obd_statfs *msfs;
2725 struct ptlrpc_request *req;
2726 struct obd_import *imp = NULL;
2730 /*Since the request might also come from lprocfs, so we need
2731 *sync this with client_disconnect_export Bug15684*/
2732 cfs_down_read(&obd->u.cli.cl_sem);
2733 if (obd->u.cli.cl_import)
2734 imp = class_import_get(obd->u.cli.cl_import);
2735 cfs_up_read(&obd->u.cli.cl_sem);
2739 /* We could possibly pass max_age in the request (as an absolute
2740 * timestamp or a "seconds.usec ago") so the target can avoid doing
2741 * extra calls into the filesystem if that isn't necessary (e.g.
2742 * during mount that would help a bit). Having relative timestamps
2743 * is not so great if request processing is slow, while absolute
2744 * timestamps are not ideal because they need time synchronization. */
2745 req = ptlrpc_request_alloc(imp, &RQF_OST_STATFS);
2747 class_import_put(imp);
2752 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_STATFS);
2754 ptlrpc_request_free(req);
2757 ptlrpc_request_set_replen(req);
2758 req->rq_request_portal = OST_CREATE_PORTAL;
2759 ptlrpc_at_set_req_timeout(req);
2761 if (flags & OBD_STATFS_NODELAY) {
2762 /* procfs requests not want stat in wait for avoid deadlock */
2763 req->rq_no_resend = 1;
2764 req->rq_no_delay = 1;
2767 rc = ptlrpc_queue_wait(req);
2771 msfs = req_capsule_server_get(&req->rq_pill, &RMF_OBD_STATFS);
2773 GOTO(out, rc = -EPROTO);
2780 ptlrpc_req_finished(req);
2784 /* Retrieve object striping information.
2786 * @lmmu is a pointer to an in-core struct with lmm_ost_count indicating
2787 * the maximum number of OST indices which will fit in the user buffer.
2788 * lmm_magic must be LOV_MAGIC (we only use 1 slot here).
2790 static int osc_getstripe(struct lov_stripe_md *lsm, struct lov_user_md *lump)
2792 /* we use lov_user_md_v3 because it is larger than lov_user_md_v1 */
2793 struct lov_user_md_v3 lum, *lumk;
2794 struct lov_user_ost_data_v1 *lmm_objects;
2795 int rc = 0, lum_size;
2801 /* we only need the header part from user space to get lmm_magic and
2802 * lmm_stripe_count, (the header part is common to v1 and v3) */
2803 lum_size = sizeof(struct lov_user_md_v1);
2804 if (cfs_copy_from_user(&lum, lump, lum_size))
2807 if ((lum.lmm_magic != LOV_USER_MAGIC_V1) &&
2808 (lum.lmm_magic != LOV_USER_MAGIC_V3))
2811 /* lov_user_md_vX and lov_mds_md_vX must have the same size */
2812 LASSERT(sizeof(struct lov_user_md_v1) == sizeof(struct lov_mds_md_v1));
2813 LASSERT(sizeof(struct lov_user_md_v3) == sizeof(struct lov_mds_md_v3));
2814 LASSERT(sizeof(lum.lmm_objects[0]) == sizeof(lumk->lmm_objects[0]));
2816 /* we can use lov_mds_md_size() to compute lum_size
2817 * because lov_user_md_vX and lov_mds_md_vX have the same size */
2818 if (lum.lmm_stripe_count > 0) {
2819 lum_size = lov_mds_md_size(lum.lmm_stripe_count, lum.lmm_magic);
2820 OBD_ALLOC(lumk, lum_size);
2824 if (lum.lmm_magic == LOV_USER_MAGIC_V1)
2825 lmm_objects = &(((struct lov_user_md_v1 *)lumk)->lmm_objects[0]);
2827 lmm_objects = &(lumk->lmm_objects[0]);
2828 lmm_objects->l_object_id = lsm->lsm_object_id;
2830 lum_size = lov_mds_md_size(0, lum.lmm_magic);
2834 lumk->lmm_object_id = lsm->lsm_object_id;
2835 lumk->lmm_object_seq = lsm->lsm_object_seq;
2836 lumk->lmm_stripe_count = 1;
2838 if (cfs_copy_to_user(lump, lumk, lum_size))
2842 OBD_FREE(lumk, lum_size);
2848 static int osc_iocontrol(unsigned int cmd, struct obd_export *exp, int len,
2849 void *karg, void *uarg)
2851 struct obd_device *obd = exp->exp_obd;
2852 struct obd_ioctl_data *data = karg;
2856 if (!cfs_try_module_get(THIS_MODULE)) {
2857 CERROR("Can't get module. Is it alive?");
2861 case OBD_IOC_LOV_GET_CONFIG: {
2863 struct lov_desc *desc;
2864 struct obd_uuid uuid;
2868 if (obd_ioctl_getdata(&buf, &len, (void *)uarg))
2869 GOTO(out, err = -EINVAL);
2871 data = (struct obd_ioctl_data *)buf;
2873 if (sizeof(*desc) > data->ioc_inllen1) {
2874 obd_ioctl_freedata(buf, len);
2875 GOTO(out, err = -EINVAL);
2878 if (data->ioc_inllen2 < sizeof(uuid)) {
2879 obd_ioctl_freedata(buf, len);
2880 GOTO(out, err = -EINVAL);
2883 desc = (struct lov_desc *)data->ioc_inlbuf1;
2884 desc->ld_tgt_count = 1;
2885 desc->ld_active_tgt_count = 1;
2886 desc->ld_default_stripe_count = 1;
2887 desc->ld_default_stripe_size = 0;
2888 desc->ld_default_stripe_offset = 0;
2889 desc->ld_pattern = 0;
2890 memcpy(&desc->ld_uuid, &obd->obd_uuid, sizeof(uuid));
2892 memcpy(data->ioc_inlbuf2, &obd->obd_uuid, sizeof(uuid));
2894 err = cfs_copy_to_user((void *)uarg, buf, len);
2897 obd_ioctl_freedata(buf, len);
2900 case LL_IOC_LOV_SETSTRIPE:
2901 err = obd_alloc_memmd(exp, karg);
2905 case LL_IOC_LOV_GETSTRIPE:
2906 err = osc_getstripe(karg, uarg);
2908 case OBD_IOC_CLIENT_RECOVER:
2909 err = ptlrpc_recover_import(obd->u.cli.cl_import,
2910 data->ioc_inlbuf1, 0);
2914 case IOC_OSC_SET_ACTIVE:
2915 err = ptlrpc_set_import_active(obd->u.cli.cl_import,
2918 case OBD_IOC_POLL_QUOTACHECK:
2919 err = osc_quota_poll_check(exp, (struct if_quotacheck *)karg);
2921 case OBD_IOC_PING_TARGET:
2922 err = ptlrpc_obd_ping(obd);
2925 CDEBUG(D_INODE, "unrecognised ioctl %#x by %s\n",
2926 cmd, cfs_curproc_comm());
2927 GOTO(out, err = -ENOTTY);
2930 cfs_module_put(THIS_MODULE);
2934 static int osc_get_info(const struct lu_env *env, struct obd_export *exp,
2935 obd_count keylen, void *key, __u32 *vallen, void *val,
2936 struct lov_stripe_md *lsm)
2939 if (!vallen || !val)
2942 if (KEY_IS(KEY_LOCK_TO_STRIPE)) {
2943 __u32 *stripe = val;
2944 *vallen = sizeof(*stripe);
2947 } else if (KEY_IS(KEY_LAST_ID)) {
2948 struct ptlrpc_request *req;
2953 req = ptlrpc_request_alloc(class_exp2cliimp(exp),
2954 &RQF_OST_GET_INFO_LAST_ID);
2958 req_capsule_set_size(&req->rq_pill, &RMF_SETINFO_KEY,
2959 RCL_CLIENT, keylen);
2960 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_GET_INFO);
2962 ptlrpc_request_free(req);
2966 tmp = req_capsule_client_get(&req->rq_pill, &RMF_SETINFO_KEY);
2967 memcpy(tmp, key, keylen);
2969 req->rq_no_delay = req->rq_no_resend = 1;
2970 ptlrpc_request_set_replen(req);
2971 rc = ptlrpc_queue_wait(req);
2975 reply = req_capsule_server_get(&req->rq_pill, &RMF_OBD_ID);
2977 GOTO(out, rc = -EPROTO);
2979 *((obd_id *)val) = *reply;
2981 ptlrpc_req_finished(req);
2983 } else if (KEY_IS(KEY_FIEMAP)) {
2984 struct ptlrpc_request *req;
2985 struct ll_user_fiemap *reply;
2989 req = ptlrpc_request_alloc(class_exp2cliimp(exp),
2990 &RQF_OST_GET_INFO_FIEMAP);
2994 req_capsule_set_size(&req->rq_pill, &RMF_FIEMAP_KEY,
2995 RCL_CLIENT, keylen);
2996 req_capsule_set_size(&req->rq_pill, &RMF_FIEMAP_VAL,
2997 RCL_CLIENT, *vallen);
2998 req_capsule_set_size(&req->rq_pill, &RMF_FIEMAP_VAL,
2999 RCL_SERVER, *vallen);
3001 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_GET_INFO);
3003 ptlrpc_request_free(req);
3007 tmp = req_capsule_client_get(&req->rq_pill, &RMF_FIEMAP_KEY);
3008 memcpy(tmp, key, keylen);
3009 tmp = req_capsule_client_get(&req->rq_pill, &RMF_FIEMAP_VAL);
3010 memcpy(tmp, val, *vallen);
3012 ptlrpc_request_set_replen(req);
3013 rc = ptlrpc_queue_wait(req);
3017 reply = req_capsule_server_get(&req->rq_pill, &RMF_FIEMAP_VAL);
3019 GOTO(out1, rc = -EPROTO);
3021 memcpy(val, reply, *vallen);
3023 ptlrpc_req_finished(req);
3031 static int osc_setinfo_mds_connect_import(struct obd_import *imp)
3033 struct llog_ctxt *ctxt;
3037 ctxt = llog_get_context(imp->imp_obd, LLOG_MDS_OST_ORIG_CTXT);
3039 rc = llog_initiator_connect(ctxt);
3040 llog_ctxt_put(ctxt);
3042 /* XXX return an error? skip setting below flags? */
3045 cfs_spin_lock(&imp->imp_lock);
3046 imp->imp_server_timeout = 1;
3047 imp->imp_pingable = 1;
3048 cfs_spin_unlock(&imp->imp_lock);
3049 CDEBUG(D_RPCTRACE, "pinging OST %s\n", obd2cli_tgt(imp->imp_obd));
3054 static int osc_setinfo_mds_conn_interpret(const struct lu_env *env,
3055 struct ptlrpc_request *req,
3062 RETURN(osc_setinfo_mds_connect_import(req->rq_import));
3065 static int osc_set_info_async(const struct lu_env *env, struct obd_export *exp,
3066 obd_count keylen, void *key, obd_count vallen,
3067 void *val, struct ptlrpc_request_set *set)
3069 struct ptlrpc_request *req;
3070 struct obd_device *obd = exp->exp_obd;
3071 struct obd_import *imp = class_exp2cliimp(exp);
3076 OBD_FAIL_TIMEOUT(OBD_FAIL_OSC_SHUTDOWN, 10);
3078 if (KEY_IS(KEY_NEXT_ID)) {
3080 struct osc_creator *oscc = &obd->u.cli.cl_oscc;
3082 if (vallen != sizeof(obd_id))
3087 if (vallen != sizeof(obd_id))
3090 /* avoid race between allocate new object and set next id
3091 * from ll_sync thread */
3092 cfs_spin_lock(&oscc->oscc_lock);
3093 new_val = *((obd_id*)val) + 1;
3094 if (new_val > oscc->oscc_next_id)
3095 oscc->oscc_next_id = new_val;
3096 cfs_spin_unlock(&oscc->oscc_lock);
3097 CDEBUG(D_HA, "%s: set oscc_next_id = "LPU64"\n",
3098 exp->exp_obd->obd_name,
3099 obd->u.cli.cl_oscc.oscc_next_id);
3104 if (KEY_IS(KEY_CHECKSUM)) {
3105 if (vallen != sizeof(int))
3107 exp->exp_obd->u.cli.cl_checksum = (*(int *)val) ? 1 : 0;
3111 if (KEY_IS(KEY_SPTLRPC_CONF)) {
3112 sptlrpc_conf_client_adapt(obd);
3116 if (KEY_IS(KEY_FLUSH_CTX)) {
3117 sptlrpc_import_flush_my_ctx(imp);
3121 if (!set && !KEY_IS(KEY_GRANT_SHRINK))
3124 /* We pass all other commands directly to OST. Since nobody calls osc
3125 methods directly and everybody is supposed to go through LOV, we
3126 assume lov checked invalid values for us.
3127 The only recognised values so far are evict_by_nid and mds_conn.
3128 Even if something bad goes through, we'd get a -EINVAL from OST
3131 if (KEY_IS(KEY_GRANT_SHRINK))
3132 req = ptlrpc_request_alloc(imp, &RQF_OST_SET_GRANT_INFO);
3134 req = ptlrpc_request_alloc(imp, &RQF_OBD_SET_INFO);
3139 req_capsule_set_size(&req->rq_pill, &RMF_SETINFO_KEY,
3140 RCL_CLIENT, keylen);
3141 req_capsule_set_size(&req->rq_pill, &RMF_SETINFO_VAL,
3142 RCL_CLIENT, vallen);
3143 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_SET_INFO);
3145 ptlrpc_request_free(req);
3149 tmp = req_capsule_client_get(&req->rq_pill, &RMF_SETINFO_KEY);
3150 memcpy(tmp, key, keylen);
3151 tmp = req_capsule_client_get(&req->rq_pill, &RMF_SETINFO_VAL);
3152 memcpy(tmp, val, vallen);
3154 if (KEY_IS(KEY_MDS_CONN)) {
3155 struct osc_creator *oscc = &obd->u.cli.cl_oscc;
3157 oscc->oscc_oa.o_seq = (*(__u32 *)val);
3158 oscc->oscc_oa.o_valid |= OBD_MD_FLGROUP;
3159 LASSERT_SEQ_IS_MDT(oscc->oscc_oa.o_seq);
3160 req->rq_no_delay = req->rq_no_resend = 1;
3161 req->rq_interpret_reply = osc_setinfo_mds_conn_interpret;
3162 } else if (KEY_IS(KEY_GRANT_SHRINK)) {
3163 struct osc_grant_args *aa;
3166 CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
3167 aa = ptlrpc_req_async_args(req);
3170 ptlrpc_req_finished(req);
3173 *oa = ((struct ost_body *)val)->oa;
3175 req->rq_interpret_reply = osc_shrink_grant_interpret;
3178 ptlrpc_request_set_replen(req);
3179 if (!KEY_IS(KEY_GRANT_SHRINK)) {
3180 LASSERT(set != NULL);
3181 ptlrpc_set_add_req(set, req);
3182 ptlrpc_check_set(NULL, set);
3184 ptlrpcd_add_req(req, PDL_POLICY_ROUND, -1);
3190 static struct llog_operations osc_size_repl_logops = {
3191 lop_cancel: llog_obd_repl_cancel
3194 static struct llog_operations osc_mds_ost_orig_logops;
3196 static int __osc_llog_init(struct obd_device *obd, struct obd_llog_group *olg,
3197 struct obd_device *tgt, struct llog_catid *catid)
3202 rc = llog_setup(obd, &obd->obd_olg, LLOG_MDS_OST_ORIG_CTXT, tgt, 1,
3203 &catid->lci_logid, &osc_mds_ost_orig_logops);
3205 CERROR("failed LLOG_MDS_OST_ORIG_CTXT\n");
3209 rc = llog_setup(obd, &obd->obd_olg, LLOG_SIZE_REPL_CTXT, tgt, 1,
3210 NULL, &osc_size_repl_logops);
3212 struct llog_ctxt *ctxt =
3213 llog_get_context(obd, LLOG_MDS_OST_ORIG_CTXT);
3216 CERROR("failed LLOG_SIZE_REPL_CTXT\n");
3221 CERROR("osc '%s' tgt '%s' catid %p rc=%d\n",
3222 obd->obd_name, tgt->obd_name, catid, rc);
3223 CERROR("logid "LPX64":0x%x\n",
3224 catid->lci_logid.lgl_oid, catid->lci_logid.lgl_ogen);
3229 static int osc_llog_init(struct obd_device *obd, struct obd_llog_group *olg,
3230 struct obd_device *disk_obd, int *index)
3232 struct llog_catid catid;
3233 static char name[32] = CATLIST;
3237 LASSERT(olg == &obd->obd_olg);
3239 cfs_mutex_lock(&olg->olg_cat_processing);
3240 rc = llog_get_cat_list(disk_obd, name, *index, 1, &catid);
3242 CERROR("rc: %d\n", rc);
3246 CDEBUG(D_INFO, "%s: Init llog for %d - catid "LPX64"/"LPX64":%x\n",
3247 obd->obd_name, *index, catid.lci_logid.lgl_oid,
3248 catid.lci_logid.lgl_oseq, catid.lci_logid.lgl_ogen);
3250 rc = __osc_llog_init(obd, olg, disk_obd, &catid);
3252 CERROR("rc: %d\n", rc);
3256 rc = llog_put_cat_list(disk_obd, name, *index, 1, &catid);
3258 CERROR("rc: %d\n", rc);
3263 cfs_mutex_unlock(&olg->olg_cat_processing);
3268 static int osc_llog_finish(struct obd_device *obd, int count)
3270 struct llog_ctxt *ctxt;
3271 int rc = 0, rc2 = 0;
3274 ctxt = llog_get_context(obd, LLOG_MDS_OST_ORIG_CTXT);
3276 rc = llog_cleanup(ctxt);
3278 ctxt = llog_get_context(obd, LLOG_SIZE_REPL_CTXT);
3280 rc2 = llog_cleanup(ctxt);
3287 static int osc_reconnect(const struct lu_env *env,
3288 struct obd_export *exp, struct obd_device *obd,
3289 struct obd_uuid *cluuid,
3290 struct obd_connect_data *data,
3293 struct client_obd *cli = &obd->u.cli;
3295 if (data != NULL && (data->ocd_connect_flags & OBD_CONNECT_GRANT)) {
3298 client_obd_list_lock(&cli->cl_loi_list_lock);
3299 data->ocd_grant = (cli->cl_avail_grant + cli->cl_dirty) ?:
3300 2 * cli->cl_max_pages_per_rpc << CFS_PAGE_SHIFT;
3301 lost_grant = cli->cl_lost_grant;
3302 cli->cl_lost_grant = 0;
3303 client_obd_list_unlock(&cli->cl_loi_list_lock);
3305 CDEBUG(D_CACHE, "request ocd_grant: %d cl_avail_grant: %ld "
3306 "cl_dirty: %ld cl_lost_grant: %ld\n", data->ocd_grant,
3307 cli->cl_avail_grant, cli->cl_dirty, lost_grant);
3308 CDEBUG(D_RPCTRACE, "ocd_connect_flags: "LPX64" ocd_version: %d"
3309 " ocd_grant: %d\n", data->ocd_connect_flags,
3310 data->ocd_version, data->ocd_grant);
3316 static int osc_disconnect(struct obd_export *exp)
3318 struct obd_device *obd = class_exp2obd(exp);
3319 struct llog_ctxt *ctxt;
3322 ctxt = llog_get_context(obd, LLOG_SIZE_REPL_CTXT);
3324 if (obd->u.cli.cl_conn_count == 1) {
3325 /* Flush any remaining cancel messages out to the
3327 llog_sync(ctxt, exp);
3329 llog_ctxt_put(ctxt);
3331 CDEBUG(D_HA, "No LLOG_SIZE_REPL_CTXT found in obd %p\n",
3335 rc = client_disconnect_export(exp);
3337 * Initially we put del_shrink_grant before disconnect_export, but it
3338 * causes the following problem if setup (connect) and cleanup
3339 * (disconnect) are tangled together.
3340 * connect p1 disconnect p2
3341 * ptlrpc_connect_import
3342 * ............... class_manual_cleanup
3345 * ptlrpc_connect_interrupt
3347 * add this client to shrink list
3349 * Bang! pinger trigger the shrink.
3350 * So the osc should be disconnected from the shrink list, after we
3351 * are sure the import has been destroyed. BUG18662
3353 if (obd->u.cli.cl_import == NULL)
3354 osc_del_shrink_grant(&obd->u.cli);
3358 static int osc_import_event(struct obd_device *obd,
3359 struct obd_import *imp,
3360 enum obd_import_event event)
3362 struct client_obd *cli;
3366 LASSERT(imp->imp_obd == obd);
3369 case IMP_EVENT_DISCON: {
3370 /* Only do this on the MDS OSC's */
3371 if (imp->imp_server_timeout) {
3372 struct osc_creator *oscc = &obd->u.cli.cl_oscc;
3374 cfs_spin_lock(&oscc->oscc_lock);
3375 oscc->oscc_flags |= OSCC_FLAG_RECOVERING;
3376 cfs_spin_unlock(&oscc->oscc_lock);
3379 client_obd_list_lock(&cli->cl_loi_list_lock);
3380 cli->cl_avail_grant = 0;
3381 cli->cl_lost_grant = 0;
3382 client_obd_list_unlock(&cli->cl_loi_list_lock);
3385 case IMP_EVENT_INACTIVE: {
3386 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_INACTIVE, NULL);
3389 case IMP_EVENT_INVALIDATE: {
3390 struct ldlm_namespace *ns = obd->obd_namespace;
3394 env = cl_env_get(&refcheck);
3398 client_obd_list_lock(&cli->cl_loi_list_lock);
3399 /* all pages go to failing rpcs due to the invalid
3401 osc_io_unplug(env, cli, NULL, PDL_POLICY_ROUND);
3402 client_obd_list_unlock(&cli->cl_loi_list_lock);
3404 ldlm_namespace_cleanup(ns, LDLM_FL_LOCAL_ONLY);
3405 cl_env_put(env, &refcheck);
3410 case IMP_EVENT_ACTIVE: {
3411 /* Only do this on the MDS OSC's */
3412 if (imp->imp_server_timeout) {
3413 struct osc_creator *oscc = &obd->u.cli.cl_oscc;
3415 cfs_spin_lock(&oscc->oscc_lock);
3416 oscc->oscc_flags &= ~(OSCC_FLAG_NOSPC |
3417 OSCC_FLAG_NOSPC_BLK);
3418 cfs_spin_unlock(&oscc->oscc_lock);
3420 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_ACTIVE, NULL);
3423 case IMP_EVENT_OCD: {
3424 struct obd_connect_data *ocd = &imp->imp_connect_data;
3426 if (ocd->ocd_connect_flags & OBD_CONNECT_GRANT)
3427 osc_init_grant(&obd->u.cli, ocd);
3430 if (ocd->ocd_connect_flags & OBD_CONNECT_REQPORTAL)
3431 imp->imp_client->cli_request_portal =OST_REQUEST_PORTAL;
3433 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_OCD, NULL);
3436 case IMP_EVENT_DEACTIVATE: {
3437 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_DEACTIVATE, NULL);
3440 case IMP_EVENT_ACTIVATE: {
3441 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_ACTIVATE, NULL);
3445 CERROR("Unknown import event %d\n", event);
3452 * Determine whether the lock can be canceled before replaying the lock
3453 * during recovery, see bug16774 for detailed information.
3455 * \retval zero the lock can't be canceled
3456 * \retval other ok to cancel
3458 static int osc_cancel_for_recovery(struct ldlm_lock *lock)
3460 check_res_locked(lock->l_resource);
3463 * Cancel all unused extent lock in granted mode LCK_PR or LCK_CR.
3465 * XXX as a future improvement, we can also cancel unused write lock
3466 * if it doesn't have dirty data and active mmaps.
3468 if (lock->l_resource->lr_type == LDLM_EXTENT &&
3469 (lock->l_granted_mode == LCK_PR ||
3470 lock->l_granted_mode == LCK_CR) &&
3471 (osc_dlm_lock_pageref(lock) == 0))
3477 static int brw_queue_work(const struct lu_env *env, void *data)
3479 struct client_obd *cli = data;
3481 CDEBUG(D_CACHE, "Run writeback work for client obd %p.\n", cli);
3483 client_obd_list_lock(&cli->cl_loi_list_lock);
3484 osc_io_unplug(env, cli, NULL, PDL_POLICY_SAME);
3485 client_obd_list_unlock(&cli->cl_loi_list_lock);
3489 int osc_setup(struct obd_device *obd, struct lustre_cfg *lcfg)
3491 struct client_obd *cli = &obd->u.cli;
3496 rc = ptlrpcd_addref();
3500 rc = client_obd_setup(obd, lcfg);
3503 handler = ptlrpcd_alloc_work(cli->cl_import,
3504 brw_queue_work, cli);
3505 if (!IS_ERR(handler))
3506 cli->cl_writeback_work = handler;
3508 rc = PTR_ERR(handler);
3512 struct lprocfs_static_vars lvars = { 0 };
3514 cli->cl_grant_shrink_interval = GRANT_SHRINK_INTERVAL;
3515 lprocfs_osc_init_vars(&lvars);
3516 if (lprocfs_obd_setup(obd, lvars.obd_vars) == 0) {
3517 lproc_osc_attach_seqstat(obd);
3518 sptlrpc_lprocfs_cliobd_attach(obd);
3519 ptlrpc_lprocfs_register_obd(obd);
3523 /* We need to allocate a few requests more, because
3524 brw_interpret tries to create new requests before freeing
3525 previous ones. Ideally we want to have 2x max_rpcs_in_flight
3526 reserved, but I afraid that might be too much wasted RAM
3527 in fact, so 2 is just my guess and still should work. */
3528 cli->cl_import->imp_rq_pool =
3529 ptlrpc_init_rq_pool(cli->cl_max_rpcs_in_flight + 2,
3531 ptlrpc_add_rqs_to_pool);
3533 CFS_INIT_LIST_HEAD(&cli->cl_grant_shrink_list);
3535 ns_register_cancel(obd->obd_namespace, osc_cancel_for_recovery);
3543 static int osc_precleanup(struct obd_device *obd, enum obd_cleanup_stage stage)
3549 case OBD_CLEANUP_EARLY: {
3550 struct obd_import *imp;
3551 imp = obd->u.cli.cl_import;
3552 CDEBUG(D_HA, "Deactivating import %s\n", obd->obd_name);
3553 /* ptlrpc_abort_inflight to stop an mds_lov_synchronize */
3554 ptlrpc_deactivate_import(imp);
3555 cfs_spin_lock(&imp->imp_lock);
3556 imp->imp_pingable = 0;
3557 cfs_spin_unlock(&imp->imp_lock);
3560 case OBD_CLEANUP_EXPORTS: {
3561 struct client_obd *cli = &obd->u.cli;
3563 * for echo client, export may be on zombie list, wait for
3564 * zombie thread to cull it, because cli.cl_import will be
3565 * cleared in client_disconnect_export():
3566 * class_export_destroy() -> obd_cleanup() ->
3567 * echo_device_free() -> echo_client_cleanup() ->
3568 * obd_disconnect() -> osc_disconnect() ->
3569 * client_disconnect_export()
3571 obd_zombie_barrier();
3572 if (cli->cl_writeback_work) {
3573 ptlrpcd_destroy_work(cli->cl_writeback_work);
3574 cli->cl_writeback_work = NULL;
3576 obd_cleanup_client_import(obd);
3577 ptlrpc_lprocfs_unregister_obd(obd);
3578 lprocfs_obd_cleanup(obd);
3579 rc = obd_llog_finish(obd, 0);
3581 CERROR("failed to cleanup llogging subsystems\n");
3588 int osc_cleanup(struct obd_device *obd)
3594 /* free memory of osc quota cache */
3595 osc_quota_cleanup(obd);
3597 rc = client_obd_cleanup(obd);
3603 int osc_process_config_base(struct obd_device *obd, struct lustre_cfg *lcfg)
3605 struct lprocfs_static_vars lvars = { 0 };
3608 lprocfs_osc_init_vars(&lvars);
3610 switch (lcfg->lcfg_command) {
3612 rc = class_process_proc_param(PARAM_OSC, lvars.obd_vars,
3622 static int osc_process_config(struct obd_device *obd, obd_count len, void *buf)
3624 return osc_process_config_base(obd, buf);
3627 struct obd_ops osc_obd_ops = {
3628 .o_owner = THIS_MODULE,
3629 .o_setup = osc_setup,
3630 .o_precleanup = osc_precleanup,
3631 .o_cleanup = osc_cleanup,
3632 .o_add_conn = client_import_add_conn,
3633 .o_del_conn = client_import_del_conn,
3634 .o_connect = client_connect_import,
3635 .o_reconnect = osc_reconnect,
3636 .o_disconnect = osc_disconnect,
3637 .o_statfs = osc_statfs,
3638 .o_statfs_async = osc_statfs_async,
3639 .o_packmd = osc_packmd,
3640 .o_unpackmd = osc_unpackmd,
3641 .o_precreate = osc_precreate,
3642 .o_create = osc_create,
3643 .o_create_async = osc_create_async,
3644 .o_destroy = osc_destroy,
3645 .o_getattr = osc_getattr,
3646 .o_getattr_async = osc_getattr_async,
3647 .o_setattr = osc_setattr,
3648 .o_setattr_async = osc_setattr_async,
3650 .o_punch = osc_punch,
3652 .o_enqueue = osc_enqueue,
3653 .o_change_cbdata = osc_change_cbdata,
3654 .o_find_cbdata = osc_find_cbdata,
3655 .o_cancel = osc_cancel,
3656 .o_cancel_unused = osc_cancel_unused,
3657 .o_iocontrol = osc_iocontrol,
3658 .o_get_info = osc_get_info,
3659 .o_set_info_async = osc_set_info_async,
3660 .o_import_event = osc_import_event,
3661 .o_llog_init = osc_llog_init,
3662 .o_llog_finish = osc_llog_finish,
3663 .o_process_config = osc_process_config,
3664 .o_quotactl = osc_quotactl,
3665 .o_quotacheck = osc_quotacheck,
3666 .o_quota_adjust_qunit = osc_quota_adjust_qunit,
3669 extern struct lu_kmem_descr osc_caches[];
3670 extern cfs_spinlock_t osc_ast_guard;
3671 extern cfs_lock_class_key_t osc_ast_guard_class;
3673 int __init osc_init(void)
3675 struct lprocfs_static_vars lvars = { 0 };
3679 /* print an address of _any_ initialized kernel symbol from this
3680 * module, to allow debugging with gdb that doesn't support data
3681 * symbols from modules.*/
3682 CDEBUG(D_INFO, "Lustre OSC module (%p).\n", &osc_caches);
3684 rc = lu_kmem_init(osc_caches);
3686 lprocfs_osc_init_vars(&lvars);
3689 rc = class_register_type(&osc_obd_ops, NULL, lvars.module_vars,
3690 LUSTRE_OSC_NAME, &osc_device_type);
3692 lu_kmem_fini(osc_caches);
3696 cfs_spin_lock_init(&osc_ast_guard);
3697 cfs_lockdep_set_class(&osc_ast_guard, &osc_ast_guard_class);
3699 osc_mds_ost_orig_logops = llog_lvfs_ops;
3700 osc_mds_ost_orig_logops.lop_setup = llog_obd_origin_setup;
3701 osc_mds_ost_orig_logops.lop_cleanup = llog_obd_origin_cleanup;
3702 osc_mds_ost_orig_logops.lop_add = llog_obd_origin_add;
3703 osc_mds_ost_orig_logops.lop_connect = llog_origin_connect;
3709 static void /*__exit*/ osc_exit(void)
3712 class_unregister_type(LUSTRE_OSC_NAME);
3713 lu_kmem_fini(osc_caches);
3716 MODULE_AUTHOR("Sun Microsystems, Inc. <http://www.lustre.org/>");
3717 MODULE_DESCRIPTION("Lustre Object Storage Client (OSC)");
3718 MODULE_LICENSE("GPL");
3720 cfs_module(osc, LUSTRE_VERSION_STRING, osc_init, osc_exit);